Skip to main content

omnigraph/exec/
mutation.rs

1use super::*;
2
3use super::query::literal_to_sql;
4
5// ─── Mutation helpers ────────────────────────────────────────────────────────
6
7/// Resolve an IRExpr to a concrete Literal value at runtime.
8fn resolve_expr_value(expr: &IRExpr, params: &ParamMap) -> Result<Literal> {
9    match expr {
10        IRExpr::Literal(lit) => Ok(lit.clone()),
11        IRExpr::Param(name) => params
12            .get(name)
13            .cloned()
14            .ok_or_else(|| OmniError::manifest(format!("parameter '{}' not provided", name))),
15        other => Err(OmniError::manifest(format!(
16            "unsupported expression in mutation: {:?}",
17            other
18        ))),
19    }
20}
21
22/// Create a single-element or N-element array from a Literal, matching the target DataType.
23fn literal_to_typed_array(
24    lit: &Literal,
25    data_type: &DataType,
26    num_rows: usize,
27) -> Result<ArrayRef> {
28    Ok(match (lit, data_type) {
29        (Literal::Null, _) => arrow_array::new_null_array(data_type, num_rows),
30        (Literal::String(s), DataType::Utf8) => {
31            Arc::new(StringArray::from(vec![s.as_str(); num_rows])) as ArrayRef
32        }
33        (Literal::Integer(n), DataType::Int32) => {
34            Arc::new(Int32Array::from(vec![*n as i32; num_rows]))
35        }
36        (Literal::Integer(n), DataType::Int64) => Arc::new(Int64Array::from(vec![*n; num_rows])),
37        (Literal::Integer(n), DataType::UInt32) => {
38            Arc::new(UInt32Array::from(vec![*n as u32; num_rows]))
39        }
40        (Literal::Integer(n), DataType::UInt64) => {
41            Arc::new(UInt64Array::from(vec![*n as u64; num_rows]))
42        }
43        (Literal::Float(f), DataType::Float32) => {
44            Arc::new(Float32Array::from(vec![*f as f32; num_rows]))
45        }
46        (Literal::Float(f), DataType::Float64) => Arc::new(Float64Array::from(vec![*f; num_rows])),
47        (Literal::Bool(b), DataType::Boolean) => Arc::new(BooleanArray::from(vec![*b; num_rows])),
48        (Literal::Date(s), DataType::Date32) => {
49            let days = crate::loader::parse_date32_literal(s)?;
50            Arc::new(Date32Array::from(vec![days; num_rows]))
51        }
52        (Literal::DateTime(s), DataType::Date64) => Arc::new(Date64Array::from(vec![
53            crate::loader::parse_date64_literal(s)?;
54            num_rows
55        ])),
56        (Literal::List(items), DataType::List(field)) => {
57            typed_list_literal_to_array(items, field.data_type(), num_rows)?
58        }
59        (Literal::List(items), DataType::FixedSizeList(field, dim))
60            if field.data_type() == &DataType::Float32 =>
61        {
62            if items.len() != *dim as usize {
63                return Err(OmniError::manifest(format!(
64                    "vector property expects {} dimensions, got {}",
65                    dim,
66                    items.len()
67                )));
68            }
69            let mut builder = FixedSizeListBuilder::with_capacity(
70                Float32Builder::with_capacity(num_rows * (*dim as usize)),
71                *dim,
72                num_rows,
73            )
74            .with_field(field.clone());
75            for _ in 0..num_rows {
76                for item in items {
77                    match item {
78                        Literal::Integer(value) => builder.values().append_value(*value as f32),
79                        Literal::Float(value) => builder.values().append_value(*value as f32),
80                        _ => {
81                            return Err(OmniError::manifest(
82                                "vector elements must be numeric".to_string(),
83                            ));
84                        }
85                    }
86                }
87                builder.append(true);
88            }
89            Arc::new(builder.finish())
90        }
91        _ => {
92            return Err(OmniError::manifest(format!(
93                "cannot convert {:?} to {:?}",
94                lit, data_type
95            )));
96        }
97    })
98}
99
100fn typed_list_literal_to_array(
101    items: &[Literal],
102    item_type: &DataType,
103    num_rows: usize,
104) -> Result<ArrayRef> {
105    match item_type {
106        DataType::Utf8 => {
107            let mut builder = ListBuilder::new(StringBuilder::new());
108            for _ in 0..num_rows {
109                for item in items {
110                    match item {
111                        Literal::String(value) => builder.values().append_value(value),
112                        _ => builder.values().append_null(),
113                    }
114                }
115                builder.append(true);
116            }
117            Ok(Arc::new(builder.finish()))
118        }
119        DataType::Boolean => {
120            let mut builder = ListBuilder::new(BooleanBuilder::new());
121            for _ in 0..num_rows {
122                for item in items {
123                    match item {
124                        Literal::Bool(value) => builder.values().append_value(*value),
125                        _ => builder.values().append_null(),
126                    }
127                }
128                builder.append(true);
129            }
130            Ok(Arc::new(builder.finish()))
131        }
132        DataType::Int32 => {
133            let mut builder = ListBuilder::new(Int32Builder::new());
134            for _ in 0..num_rows {
135                for item in items {
136                    match item {
137                        Literal::Integer(value) => {
138                            let value = i32::try_from(*value).map_err(|_| {
139                                OmniError::manifest(format!(
140                                    "list value {} exceeds Int32 range",
141                                    value
142                                ))
143                            })?;
144                            builder.values().append_value(value);
145                        }
146                        _ => builder.values().append_null(),
147                    }
148                }
149                builder.append(true);
150            }
151            Ok(Arc::new(builder.finish()))
152        }
153        DataType::Int64 => {
154            let mut builder = ListBuilder::new(Int64Builder::new());
155            for _ in 0..num_rows {
156                for item in items {
157                    match item {
158                        Literal::Integer(value) => builder.values().append_value(*value),
159                        _ => builder.values().append_null(),
160                    }
161                }
162                builder.append(true);
163            }
164            Ok(Arc::new(builder.finish()))
165        }
166        DataType::UInt32 => {
167            let mut builder = ListBuilder::new(UInt32Builder::new());
168            for _ in 0..num_rows {
169                for item in items {
170                    match item {
171                        Literal::Integer(value) => {
172                            let value = u32::try_from(*value).map_err(|_| {
173                                OmniError::manifest(format!(
174                                    "list value {} exceeds UInt32 range",
175                                    value
176                                ))
177                            })?;
178                            builder.values().append_value(value);
179                        }
180                        _ => builder.values().append_null(),
181                    }
182                }
183                builder.append(true);
184            }
185            Ok(Arc::new(builder.finish()))
186        }
187        DataType::UInt64 => {
188            let mut builder = ListBuilder::new(UInt64Builder::new());
189            for _ in 0..num_rows {
190                for item in items {
191                    match item {
192                        Literal::Integer(value) => {
193                            let value = u64::try_from(*value).map_err(|_| {
194                                OmniError::manifest(format!(
195                                    "list value {} exceeds UInt64 range",
196                                    value
197                                ))
198                            })?;
199                            builder.values().append_value(value);
200                        }
201                        _ => builder.values().append_null(),
202                    }
203                }
204                builder.append(true);
205            }
206            Ok(Arc::new(builder.finish()))
207        }
208        DataType::Float32 => {
209            let mut builder = ListBuilder::new(Float32Builder::new());
210            for _ in 0..num_rows {
211                for item in items {
212                    match item {
213                        Literal::Integer(value) => builder.values().append_value(*value as f32),
214                        Literal::Float(value) => builder.values().append_value(*value as f32),
215                        _ => builder.values().append_null(),
216                    }
217                }
218                builder.append(true);
219            }
220            Ok(Arc::new(builder.finish()))
221        }
222        DataType::Float64 => {
223            let mut builder = ListBuilder::new(Float64Builder::new());
224            for _ in 0..num_rows {
225                for item in items {
226                    match item {
227                        Literal::Integer(value) => builder.values().append_value(*value as f64),
228                        Literal::Float(value) => builder.values().append_value(*value),
229                        _ => builder.values().append_null(),
230                    }
231                }
232                builder.append(true);
233            }
234            Ok(Arc::new(builder.finish()))
235        }
236        DataType::Date32 => {
237            let mut builder = ListBuilder::new(Date32Builder::new());
238            for _ in 0..num_rows {
239                for item in items {
240                    match item {
241                        Literal::Date(value) => builder
242                            .values()
243                            .append_value(crate::loader::parse_date32_literal(value)?),
244                        _ => builder.values().append_null(),
245                    }
246                }
247                builder.append(true);
248            }
249            Ok(Arc::new(builder.finish()))
250        }
251        DataType::Date64 => {
252            let mut builder = ListBuilder::new(Date64Builder::new());
253            for _ in 0..num_rows {
254                for item in items {
255                    match item {
256                        Literal::DateTime(value) => builder
257                            .values()
258                            .append_value(crate::loader::parse_date64_literal(value)?),
259                        _ => builder.values().append_null(),
260                    }
261                }
262                builder.append(true);
263            }
264            Ok(Arc::new(builder.finish()))
265        }
266        other => Err(OmniError::manifest(format!(
267            "cannot convert list literal to {:?}",
268            other
269        ))),
270    }
271}
272
273/// Build a single-element blob array from a URI or base64 value string.
274fn build_blob_array_from_value(value: &str) -> Result<ArrayRef> {
275    let mut builder = BlobArrayBuilder::new(1);
276    crate::loader::append_blob_value(&mut builder, value)?;
277    builder
278        .finish()
279        .map_err(|e| OmniError::Lance(e.to_string()))
280}
281
282/// Build a null blob array with one element.
283fn build_null_blob_array() -> Result<ArrayRef> {
284    let mut builder = BlobArrayBuilder::new(1);
285    builder
286        .push_null()
287        .map_err(|e| OmniError::Lance(e.to_string()))?;
288    builder
289        .finish()
290        .map_err(|e| OmniError::Lance(e.to_string()))
291}
292
293/// Build a single-row RecordBatch from resolved assignments.
294fn build_insert_batch(
295    schema: &SchemaRef,
296    id: &str,
297    assignments: &HashMap<String, Literal>,
298    blob_properties: &HashSet<String>,
299) -> Result<RecordBatch> {
300    let mut columns: Vec<ArrayRef> = Vec::with_capacity(schema.fields().len());
301
302    for field in schema.fields() {
303        if field.name() == "id" {
304            columns.push(Arc::new(StringArray::from(vec![id])));
305        } else if blob_properties.contains(field.name()) {
306            if let Some(Literal::String(uri)) = assignments.get(field.name()) {
307                columns.push(build_blob_array_from_value(uri)?);
308            } else if field.is_nullable() {
309                columns.push(build_null_blob_array()?);
310            } else {
311                return Err(OmniError::manifest(format!(
312                    "missing required blob property '{}'",
313                    field.name()
314                )));
315            }
316        } else if field.name() == "src" {
317            let lit = assignments.get("from").ok_or_else(|| {
318                OmniError::manifest("missing required edge endpoint 'from'".to_string())
319            })?;
320            columns.push(literal_to_typed_array(lit, field.data_type(), 1)?);
321        } else if field.name() == "dst" {
322            let lit = assignments.get("to").ok_or_else(|| {
323                OmniError::manifest("missing required edge endpoint 'to'".to_string())
324            })?;
325            columns.push(literal_to_typed_array(lit, field.data_type(), 1)?);
326        } else if let Some(lit) = assignments.get(field.name()) {
327            columns.push(literal_to_typed_array(lit, field.data_type(), 1)?);
328        } else if field.is_nullable() {
329            columns.push(arrow_array::new_null_array(field.data_type(), 1));
330        } else {
331            return Err(OmniError::manifest(format!(
332                "missing required property '{}'",
333                field.name()
334            )));
335        }
336    }
337
338    RecordBatch::try_new(schema.clone(), columns).map_err(|e| OmniError::Lance(e.to_string()))
339}
340
341async fn validate_edge_insert_endpoints(
342    db: &Omnigraph,
343    staging: &MutationStaging,
344    branch: Option<&str>,
345    edge_name: &str,
346    assignments: &HashMap<String, Literal>,
347) -> Result<()> {
348    let catalog = db.catalog();
349    let edge_type = catalog
350        .edge_types
351        .get(edge_name)
352        .ok_or_else(|| OmniError::manifest(format!("unknown edge type '{}'", edge_name)))?;
353    let from = match assignments.get("from") {
354        Some(Literal::String(value)) => value.as_str(),
355        Some(other) => {
356            return Err(OmniError::manifest(format!(
357                "edge {} from endpoint must be a string id, got {}",
358                edge_name,
359                literal_to_sql(other)
360            )));
361        }
362        None => {
363            return Err(OmniError::manifest(format!(
364                "edge {} missing 'from' endpoint",
365                edge_name
366            )));
367        }
368    };
369    let to = match assignments.get("to") {
370        Some(Literal::String(value)) => value.as_str(),
371        Some(other) => {
372            return Err(OmniError::manifest(format!(
373                "edge {} to endpoint must be a string id, got {}",
374                edge_name,
375                literal_to_sql(other)
376            )));
377        }
378        None => {
379            return Err(OmniError::manifest(format!(
380                "edge {} missing 'to' endpoint",
381                edge_name
382            )));
383        }
384    };
385
386    ensure_node_id_exists(db, staging, branch, &edge_type.from_type, from, "src").await?;
387    ensure_node_id_exists(db, staging, branch, &edge_type.to_type, to, "dst").await?;
388    Ok(())
389}
390
391/// Quick scan of pending batches for an `id` value match. Used by the
392/// mutation path's edge endpoint validation to satisfy read-your-writes
393/// for same-query inserts before they're committed to Lance.
394fn pending_batches_contain_id(batches: &[RecordBatch], id: &str) -> bool {
395    for batch in batches {
396        let Some(col) = batch.column_by_name("id") else {
397            continue;
398        };
399        let Some(arr) = col.as_any().downcast_ref::<StringArray>() else {
400            continue;
401        };
402        for i in 0..arr.len() {
403            if arr.is_valid(i) && arr.value(i) == id {
404                return true;
405            }
406        }
407    }
408    false
409}
410
411async fn ensure_node_id_exists(
412    db: &Omnigraph,
413    staging: &MutationStaging,
414    branch: Option<&str>,
415    node_type: &str,
416    id: &str,
417    label: &str,
418) -> Result<()> {
419    let table_key = format!("node:{}", node_type);
420
421    // Prefer the in-query pending accumulator so a same-query insert of
422    // the referenced node is visible to this validation. Fall back to
423    // the pre-mutation manifest snapshot when nothing pending matches.
424    let pending = staging.pending_batches(&table_key);
425    if pending_batches_contain_id(pending, id) {
426        return Ok(());
427    }
428
429    let filter = format!("id = '{}'", id.replace('\'', "''"));
430    let snapshot = db.snapshot_for_branch(branch).await?;
431    let ds = db
432        .storage()
433        .open_snapshot_at_table(&snapshot, &table_key)
434        .await?;
435    let exists = db.storage().count_rows(&ds, Some(filter)).await? > 0;
436
437    if exists {
438        Ok(())
439    } else {
440        Err(OmniError::manifest(format!(
441            "{} '{}' not found in {}",
442            label, id, node_type
443        )))
444    }
445}
446
447/// Convert an IRMutationPredicate to a Lance SQL filter string.
448fn predicate_to_sql(
449    predicate: &IRMutationPredicate,
450    params: &ParamMap,
451    is_edge: bool,
452) -> Result<String> {
453    let column = if is_edge {
454        match predicate.property.as_str() {
455            "from" => "src".to_string(),
456            "to" => "dst".to_string(),
457            other => other.to_string(),
458        }
459    } else {
460        predicate.property.clone()
461    };
462
463    let value = resolve_expr_value(&predicate.value, params)?;
464    let value_sql = literal_to_sql(&value);
465
466    let op = match predicate.op {
467        CompOp::Eq => "=",
468        CompOp::Ne => "!=",
469        CompOp::Gt => ">",
470        CompOp::Lt => "<",
471        CompOp::Ge => ">=",
472        CompOp::Le => "<=",
473        CompOp::Contains => {
474            return Err(OmniError::manifest(
475                "contains predicate not supported in mutations".to_string(),
476            ));
477        }
478    };
479
480    Ok(format!("{} {} {}", column, op, value_sql))
481}
482
483/// Replace specific columns in a RecordBatch with new literal values.
484///
485/// Blob columns may or may not be present in `batch` depending on the
486/// caller's scan projection:
487/// - If `batch` does NOT contain a blob column AND it has no assignment,
488///   the column is OMITTED from the output. `merge_insert` leaves it
489///   untouched.
490/// - If `batch` DOES contain a blob column AND it has no assignment, the
491///   column is COPIED to the output. This enables coalescing of
492///   different-shape updates into a single full-schema merge batch (the
493///   per-table accumulator in `MutationStaging` requires consistent
494///   schemas across pending batches for `concat_batches`). The
495///   round-tripping cost is acceptable for typical agent-driven
496///   mutations; tables with large blobs and unassigned-blob updates may
497///   want to be split into separate queries.
498/// - If a blob column has a string-URI assignment, build the blob array
499///   inline.
500fn apply_assignments(
501    full_schema: &SchemaRef,
502    batch: &RecordBatch,
503    assignments: &HashMap<String, Literal>,
504    blob_properties: &HashSet<String>,
505) -> Result<RecordBatch> {
506    let mut columns: Vec<ArrayRef> = Vec::with_capacity(full_schema.fields().len());
507    let mut out_fields: Vec<Field> = Vec::with_capacity(full_schema.fields().len());
508
509    for field in full_schema.fields().iter() {
510        if blob_properties.contains(field.name()) {
511            if let Some(Literal::String(uri)) = assignments.get(field.name()) {
512                // Assigned: build a single blob column from the URI.
513                let mut builder = BlobArrayBuilder::new(batch.num_rows());
514                for _ in 0..batch.num_rows() {
515                    crate::loader::append_blob_value(&mut builder, uri)?;
516                }
517                let blob_field = lance::blob::blob_field(field.name(), true);
518                out_fields.push(blob_field);
519                columns.push(
520                    builder
521                        .finish()
522                        .map_err(|e| OmniError::Lance(e.to_string()))?,
523                );
524            } else if let Some(col) = batch.column_by_name(field.name()) {
525                // Unassigned but scan included it: copy through (writes
526                // back the same blob, no observable change but uniform
527                // schema for the accumulator).
528                let blob_field = lance::blob::blob_field(field.name(), field.is_nullable());
529                out_fields.push(blob_field);
530                columns.push(col.clone());
531            }
532            // else: scan did not include this blob column and no
533            // assignment — omit. Caller's accumulator must accept the
534            // narrower schema (legacy single-merge_insert path).
535        } else if let Some(lit) = assignments.get(field.name()) {
536            out_fields.push(field.as_ref().clone());
537            columns.push(literal_to_typed_array(
538                lit,
539                field.data_type(),
540                batch.num_rows(),
541            )?);
542        } else {
543            let col = batch.column_by_name(field.name()).ok_or_else(|| {
544                OmniError::Lance(format!(
545                    "column '{}' not found in scan result",
546                    field.name()
547                ))
548            })?;
549            out_fields.push(field.as_ref().clone());
550            columns.push(col.clone());
551        }
552    }
553
554    RecordBatch::try_new(Arc::new(Schema::new(out_fields)), columns)
555        .map_err(|e| OmniError::Lance(e.to_string()))
556}
557
558// ─── Mutation execution ──────────────────────────────────────────────────────
559
560use super::staging::{MutationStaging, PendingMode};
561
562/// Open a sub-table dataset for read or inline-commit-write within the
563/// current mutation query, capturing pre-write metadata in `staging` on
564/// first touch. The captured version is the publisher's CAS fence at
565/// end-of-query (per-table OCC).
566///
567/// On first touch, opens the dataset at HEAD on the requested branch
568/// via `open_for_mutation_on_branch`, which compares Lance HEAD against
569/// the manifest's pinned version — that fence is the engine's
570/// publisher-style OCC catching cross-writer drift before we make any
571/// changes. For delete-only queries, this strict open is also the uncovered
572/// drift guard that runs before `delete_where` can inline-commit.
573///
574/// On subsequent touches *within the same query*, behavior depends on
575/// whether the table has already been inline-committed by a delete op:
576///
577/// - **Insert / update path (no inline commit between touches).** Lance
578///   HEAD has not moved since first touch, so a fresh
579///   `open_for_mutation_on_branch` would still match the manifest
580///   pinned version. We just go through it again; `ensure_path` is a
581///   no-op (idempotent on the captured `expected_version`).
582/// - **Delete cascade or multi-delete on the same table.** A prior
583///   `delete_where` on this table has already advanced Lance HEAD past
584///   the manifest's pinned version (the manifest doesn't move until
585///   end-of-query). Going through `open_for_mutation_on_branch` again
586///   would trip its `ensure_expected_version` equality check
587///   (`actual = pinned + 1` vs `expected = pinned`). Instead we route
588///   through `reopen_for_mutation` at the post-inline-commit Lance
589///   version captured in `staging.inline_committed[table_key]`, which
590///   is the source of truth for "where is Lance HEAD right now on
591///   this table within this query."
592///
593/// The `inline_committed` reopen branch closes the multi-delete-on-same-table
594/// failure path that pre-staged-write engines inherited. The branch goes
595/// away once Lance exposes a two-phase delete API
596/// ([lance-format/lance#6658](https://github.com/lance-format/lance/issues/6658))
597/// and we can stage deletes on the same path as inserts/updates.
598async fn open_table_for_mutation(
599    db: &Omnigraph,
600    staging: &mut MutationStaging,
601    branch: Option<&str>,
602    table_key: &str,
603    op_kind: crate::db::MutationOpKind,
604) -> Result<(SnapshotHandle, String, Option<String>)> {
605    if let Some(prior) = staging.inline_committed.get(table_key) {
606        let path = staging.paths.get(table_key).ok_or_else(|| {
607            OmniError::manifest_internal(format!(
608                "open_table_for_mutation: inline_committed[{}] without paths entry",
609                table_key
610            ))
611        })?;
612        let ds = db
613            .reopen_for_mutation(
614                table_key,
615                &path.full_path,
616                path.table_branch.as_deref(),
617                prior.table_version,
618                op_kind,
619            )
620            .await?;
621        return Ok((ds, path.full_path.clone(), path.table_branch.clone()));
622    }
623    let (ds, full_path, table_branch) = db
624        .open_for_mutation_on_branch(branch, table_key, op_kind)
625        .await?;
626    let expected_version = ds.version();
627    staging.ensure_path(
628        table_key,
629        full_path.clone(),
630        table_branch.clone(),
631        expected_version,
632        op_kind,
633    );
634    Ok((ds, full_path, table_branch))
635}
636
637/// D₂ parse-time check: a single mutation query is either insert/update-only
638/// or delete-only. Mixed → reject before any I/O.
639///
640/// Reason: under the staged-write writer, inserts and updates
641/// accumulate in memory and commit at end-of-query, while deletes still
642/// inline-commit (Lance lacks a public two-phase delete in 6.0.1).
643/// Mixing creates ordering hazards (same-row insert→delete becomes a no-op
644/// because the staged insert isn't visible to delete; cascading deletes
645/// of just-inserted edges break referential integrity by silent design).
646/// Until Lance exposes `DeleteJob::execute_uncommitted`, the parse-time
647/// rejection keeps both paths atomic and correct.
648fn enforce_no_mixed_destructive_constructive(
649    ir: &omnigraph_compiler::ir::MutationIR,
650) -> Result<()> {
651    let mut has_constructive = false;
652    let mut has_delete = false;
653    for op in &ir.ops {
654        match op {
655            MutationOpIR::Insert { .. } | MutationOpIR::Update { .. } => {
656                has_constructive = true;
657            }
658            MutationOpIR::Delete { .. } => {
659                has_delete = true;
660            }
661        }
662    }
663    if has_constructive && has_delete {
664        return Err(OmniError::manifest(format!(
665            "mutation '{}' on the same query mixes inserts/updates and deletes; \
666             split into separate mutations: (1) inserts and updates, then (2) deletes. \
667             This restriction lifts when Lance exposes a two-phase delete API \
668             (tracked: lance-format/lance#6658).",
669            ir.name
670        )));
671    }
672    Ok(())
673}
674
675impl Omnigraph {
676    pub async fn mutate(
677        &self,
678        branch: &str,
679        query_source: &str,
680        query_name: &str,
681        params: &ParamMap,
682    ) -> Result<MutationResult> {
683        self.mutate_as(branch, query_source, query_name, params, None)
684            .await
685    }
686
687    pub async fn mutate_as(
688        &self,
689        branch: &str,
690        query_source: &str,
691        query_name: &str,
692        params: &ParamMap,
693        actor_id: Option<&str>,
694    ) -> Result<MutationResult> {
695        // Engine-layer policy gate (MR-722 fan-out / PR #3). Scope is
696        // `Branch(branch)` to match the HTTP-layer convention at
697        // `server_change` (branch=Some(branch), target_branch=None). When no
698        // PolicyChecker is installed this is a no-op; with policy installed
699        // and actor=None this fails hard (forget-the-actor footgun guard).
700        self.enforce(
701            omnigraph_policy::PolicyAction::Change,
702            &omnigraph_policy::ResourceScope::Branch(branch.to_string()),
703            actor_id,
704        )?;
705        self.mutate_with_current_actor(branch, query_source, query_name, params, actor_id)
706            .await
707    }
708
709    async fn mutate_with_current_actor(
710        &self,
711        branch: &str,
712        query_source: &str,
713        query_name: &str,
714        params: &ParamMap,
715        actor_id: Option<&str>,
716    ) -> Result<MutationResult> {
717        self.ensure_schema_state_valid().await?;
718        // Converge any pending recovery sidecar (a previously failed
719        // writer's Phase B → Phase C residual) before executing: the
720        // inline delete path advances Lance HEAD during execution and
721        // the staged path's commit-time drift guard refuses
722        // sidecar-covered drift, so a long-lived handle must heal here
723        // — not at restart. One `list_dir` when no sidecars exist (the
724        // steady state).
725        self.heal_pending_recovery_sidecars().await?;
726        let requested = Self::normalize_branch_name(branch)?;
727        // Reject internal `__run__*` / system-prefixed branches at the
728        // public write boundary. Direct-publish paths assert this
729        // explicitly so a caller can't write to legacy or system
730        // staging branches by passing the prefix verbatim.
731        if let Some(name) = requested.as_deref() {
732            crate::db::ensure_public_branch_ref(name, "mutate")?;
733        }
734        let resolved_params = enrich_mutation_params(params)?;
735
736        // Per-query staging accumulator. Inserts and updates push batches
737        // into `pending`; deletes still inline-commit and record into
738        // `inline_committed`. At end-of-query, `finalize` issues one
739        // `stage_*` + `commit_staged` per pending table, then the
740        // publisher commits the manifest atomically across all touched
741        // tables. Branch is threaded explicitly — no coordinator swap.
742        let mut staging = MutationStaging::default();
743
744        // Lower + validate up front so the touched-table set is known before
745        // execution. A lowering/validation error returns exactly as it did
746        // when this happened inside execute_named_mutation.
747        let ir = self.lower_named_mutation(query_source, query_name)?;
748
749        // Up-front fork-queue acquisition (see the loader for the full
750        // rationale): if this mutation will fork any touched table onto a
751        // non-main branch, acquire the per-(table, branch) write queues for
752        // every touched table before the first fork and hold them through the
753        // publish, so the orphan-fork reclaim can't race a concurrent
754        // in-process fork. The touched set is derived from the lowered IR.
755        let fork_queue_guards: Option<(
756            Vec<(String, Option<String>)>,
757            Vec<tokio::sync::OwnedMutexGuard<()>>,
758        )> = if let Some(active) = requested.as_deref() {
759            let snapshot = self.snapshot_for_branch(Some(active)).await?;
760            let touched: Vec<(String, Option<String>)> = self
761                .touched_table_keys(&ir)
762                .into_iter()
763                .map(|k| (k, Some(active.to_string())))
764                .collect();
765            let needs_fork = touched.iter().any(|(table_key, _)| {
766                snapshot
767                    .entry(table_key)
768                    .map(|e| e.table_branch.as_deref() != Some(active))
769                    .unwrap_or(false)
770            });
771            if needs_fork {
772                let guards = self.write_queue().acquire_many(&touched).await;
773                Some((touched, guards))
774            } else {
775                None
776            }
777        } else {
778            None
779        };
780
781        let exec_result = self
782            .execute_named_mutation(&ir, &resolved_params, requested.as_deref(), &mut staging)
783            .await;
784
785        match exec_result {
786            Err(e) => Err(e),
787            Ok(total) if staging.is_empty() => Ok(total),
788            Ok(total) => {
789                let staged = staging.stage_all(self, requested.as_deref()).await?;
790                // `_queue_guards` holds per-(table_key, branch) write
791                // queues acquired inside `commit_all`. Held across the
792                // manifest publish below so no concurrent writer can
793                // interleave between our commit_staged and our publish
794                // (which would correctly fail our CAS but leave Lance
795                // HEAD advanced — the residual class MR-870 recovers).
796                let (updates, expected_versions, sidecar_handle, _queue_guards) = staged
797                    .commit_all(
798                        self,
799                        requested.as_deref(),
800                        crate::db::manifest::SidecarKind::Mutation,
801                        actor_id,
802                        fork_queue_guards,
803                    )
804                    .await?;
805                // Failpoint that wedges the documented finalize→publisher
806                // residual: per-table `commit_staged` calls already
807                // advanced Lance HEAD on every touched table; a failure
808                // injected here mirrors the production-rare case where
809                // the publisher's CAS pre-check rejects (or the manifest
810                // write throws) after staged commits succeeded. The
811                // sidecar written inside `staging.finalize()` persists
812                // across this failure so the next `Omnigraph::open`'s
813                // recovery sweep can roll forward — see
814                // `tests/failpoints.rs::recovery_rolls_forward_after_finalize_publisher_failure`.
815                crate::failpoints::maybe_fail("mutation.post_finalize_pre_publisher")?;
816                self.commit_updates_on_branch_with_expected(
817                    requested.as_deref(),
818                    &updates,
819                    &expected_versions,
820                    actor_id,
821                )
822                .await?;
823                // Phase C succeeded — sidecar can be deleted. If this
824                // delete fails, the next open's sweep classifies every
825                // table as NoMovement (manifest pin == Lance HEAD ==
826                // post_commit_pin) and the sidecar is treated as a
827                // stale artifact (cleaned up via the Phase 2 logic).
828                if let Some(handle) = sidecar_handle {
829                    // Best-effort cleanup: the manifest publish already
830                    // succeeded, so the user's mutation is durable. A
831                    // failed delete leaves the sidecar on disk; the
832                    // next open's recovery sweep classifies every table
833                    // as `NoMovement` (manifest pin == Lance HEAD ==
834                    // post_commit_pin) and tidies up. Failing the user
835                    // here would return an error for a write that
836                    // already landed.
837                    if let Err(err) =
838                        crate::db::manifest::delete_sidecar(&handle, self.storage_adapter()).await
839                    {
840                        tracing::warn!(
841                            error = %err,
842                            operation_id = handle.operation_id.as_str(),
843                            "recovery sidecar cleanup failed; the next open's recovery sweep will resolve it"
844                        );
845                    }
846                }
847                Ok(total)
848            }
849        }
850    }
851
852    /// Lower + validate a named mutation query into its IR.
853    ///
854    /// Hoisted out of [`Self::execute_named_mutation`] so the caller can
855    /// inspect the IR before execution — specifically to compute the
856    /// touched-table set (see [`Self::touched_table_keys`]) for up-front
857    /// write-queue acquisition. Performs the same find → typecheck → lower
858    /// → D₂ checks that execution previously did inline, so error behavior
859    /// is unchanged.
860    fn lower_named_mutation(
861        &self,
862        query_source: &str,
863        query_name: &str,
864    ) -> Result<omnigraph_compiler::ir::MutationIR> {
865        let query_decl = omnigraph_compiler::find_named_query(query_source, query_name)
866            .map_err(|e| OmniError::manifest(e.to_string()))?;
867
868        let checked = typecheck_query_decl(&self.catalog(), &query_decl)?;
869        match checked {
870            CheckedQuery::Mutation(_) => {}
871            CheckedQuery::Read(_) => {
872                return Err(OmniError::manifest(
873                    "mutation execution called on a read query; use query instead".to_string(),
874                ));
875            }
876        }
877
878        let ir = lower_mutation_query(&query_decl)?;
879        // D₂: reject mixed insert/update + delete before any I/O.
880        enforce_no_mixed_destructive_constructive(&ir)?;
881        Ok(ir)
882    }
883
884    /// The COMPLETE set of `(node|edge):{type}` table keys a mutation IR can
885    /// touch at execution time, keyed as `MutationStaging`/`commit_all` key
886    /// them. Must be a superset of everything execution forks/commits, since
887    /// it drives the up-front fork-queue acquisition and `commit_all`'s
888    /// held-guard coverage check — a miss means an unserialized fork/commit.
889    ///
890    /// The set is a pure function of (IR ops + catalog). For each op it mirrors
891    /// the execute path's node-vs-edge dispatch (`node_types` first, then
892    /// `edge_types`). A `delete <Node>` additionally **cascades** to every edge
893    /// type whose endpoint is that node (see `execute_delete_node`), forking
894    /// those edge tables during execution — so they are included here, derived
895    /// the same way the executor derives them (`from_type`/`to_type` match).
896    /// Unknown types are skipped (the execute path surfaces the error).
897    /// Sorted + deduped for one-shot `acquire_many`.
898    fn touched_table_keys(&self, ir: &omnigraph_compiler::ir::MutationIR) -> Vec<String> {
899        use omnigraph_compiler::ir::MutationOpIR;
900        let catalog = self.catalog();
901        let mut keys: Vec<String> = Vec::new();
902        for op in &ir.ops {
903            let type_name = match op {
904                MutationOpIR::Insert { type_name, .. }
905                | MutationOpIR::Update { type_name, .. }
906                | MutationOpIR::Delete { type_name, .. } => type_name,
907            };
908            if catalog.node_types.contains_key(type_name) {
909                keys.push(format!("node:{type_name}"));
910                // A node delete cascades to every edge touching this node type,
911                // forking those edge tables. Include them so the up-front
912                // acquisition covers the cascade (mirrors execute_delete_node).
913                if matches!(op, MutationOpIR::Delete { .. }) {
914                    for (edge_name, edge_type) in &catalog.edge_types {
915                        if edge_type.from_type == *type_name || edge_type.to_type == *type_name {
916                            keys.push(format!("edge:{edge_name}"));
917                        }
918                    }
919                }
920            } else if catalog.edge_types.contains_key(type_name) {
921                keys.push(format!("edge:{type_name}"));
922            }
923        }
924        keys.sort();
925        keys.dedup();
926        keys
927    }
928
929    async fn execute_named_mutation(
930        &self,
931        ir: &omnigraph_compiler::ir::MutationIR,
932        params: &ParamMap,
933        branch: Option<&str>,
934        staging: &mut MutationStaging,
935    ) -> Result<MutationResult> {
936        let mut total = MutationResult::default();
937        for op in &ir.ops {
938            let result = match op {
939                MutationOpIR::Insert {
940                    type_name,
941                    assignments,
942                } => {
943                    self.execute_insert(type_name, assignments, params, branch, staging)
944                        .await?
945                }
946                MutationOpIR::Update {
947                    type_name,
948                    assignments,
949                    predicate,
950                } => {
951                    self.execute_update(type_name, assignments, predicate, params, branch, staging)
952                        .await?
953                }
954                MutationOpIR::Delete {
955                    type_name,
956                    predicate,
957                } => {
958                    self.execute_delete(type_name, predicate, params, branch, staging)
959                        .await?
960                }
961            };
962            total.affected_nodes += result.affected_nodes;
963            total.affected_edges += result.affected_edges;
964        }
965        Ok(total)
966    }
967
968    async fn execute_insert(
969        &self,
970        type_name: &str,
971        assignments: &[IRAssignment],
972        params: &ParamMap,
973        branch: Option<&str>,
974        staging: &mut MutationStaging,
975    ) -> Result<MutationResult> {
976        let mut resolved: HashMap<String, Literal> = HashMap::new();
977        for a in assignments {
978            resolved.insert(a.property.clone(), resolve_expr_value(&a.value, params)?);
979        }
980
981        let is_node = self.catalog().node_types.contains_key(type_name);
982        let is_edge = self.catalog().edge_types.contains_key(type_name);
983
984        if is_node {
985            let node_type = &self.catalog().node_types[type_name];
986            let schema = node_type.arrow_schema.clone();
987            let blob_props = node_type.blob_properties.clone();
988            let id = if let Some(key_prop) = node_type.key_property() {
989                match resolved.get(key_prop) {
990                    Some(Literal::String(s)) => s.clone(),
991                    Some(other) => literal_to_sql(other).trim_matches('\'').to_string(),
992                    None => {
993                        return Err(OmniError::manifest(format!(
994                            "insert missing @key property '{}'",
995                            key_prop
996                        )));
997                    }
998                }
999            } else {
1000                ulid::Ulid::new().to_string()
1001            };
1002
1003            let batch = build_insert_batch(&schema, &id, &resolved, &blob_props)?;
1004            crate::loader::validate_value_constraints(&batch, node_type)?;
1005            crate::loader::validate_enum_constraints(&batch, &node_type.properties, type_name)?;
1006            let unique_groups = crate::loader::unique_constraint_groups_for_node(node_type);
1007            if !unique_groups.is_empty() {
1008                crate::loader::enforce_unique_constraints_intra_batch(
1009                    &batch,
1010                    type_name,
1011                    &unique_groups,
1012                )?;
1013            }
1014            let has_key = node_type.key_property().is_some();
1015            let table_key = format!("node:{}", type_name);
1016            // Capture pre-write metadata on first touch (no Lance write).
1017            let insert_kind = if has_key {
1018                crate::db::MutationOpKind::Merge
1019            } else {
1020                crate::db::MutationOpKind::Insert
1021            };
1022            let (_ds, _full_path, _table_branch) =
1023                open_table_for_mutation(self, staging, branch, &table_key, insert_kind).await?;
1024            // Accumulate. @key inserts go into the Merge stream (so a
1025            // later update on the same id coalesces correctly); no-key
1026            // inserts go into the Append stream.
1027            let mode = if has_key {
1028                PendingMode::Merge
1029            } else {
1030                PendingMode::Append
1031            };
1032            staging.append_batch(&table_key, schema, mode, batch)?;
1033
1034            Ok(MutationResult {
1035                affected_nodes: 1,
1036                affected_edges: 0,
1037            })
1038        } else if is_edge {
1039            let edge_type = &self.catalog().edge_types[type_name];
1040            let schema = edge_type.arrow_schema.clone();
1041            let blob_props = edge_type.blob_properties.clone();
1042            let id = ulid::Ulid::new().to_string();
1043
1044            let batch = build_insert_batch(&schema, &id, &resolved, &blob_props)?;
1045            validate_edge_insert_endpoints(self, staging, branch, type_name, &resolved).await?;
1046            crate::loader::validate_enum_constraints(&batch, &edge_type.properties, type_name)?;
1047            let unique_groups = crate::loader::unique_constraint_groups_for_edge(edge_type);
1048            if !unique_groups.is_empty() {
1049                crate::loader::enforce_unique_constraints_intra_batch(
1050                    &batch,
1051                    type_name,
1052                    &unique_groups,
1053                )?;
1054            }
1055            let table_key = format!("edge:{}", type_name);
1056            // Capture pre-write metadata on first touch (no Lance write).
1057            let (ds, _full_path, _table_branch) = open_table_for_mutation(
1058                self,
1059                staging,
1060                branch,
1061                &table_key,
1062                crate::db::MutationOpKind::Insert,
1063            )
1064            .await?;
1065            // Accumulate the new edge row. Edge IDs are ULID-generated so
1066            // Append mode is correct (no key-based dedup needed).
1067            staging.append_batch(&table_key, schema, PendingMode::Append, batch.clone())?;
1068
1069            // Edge cardinality validation: scan committed edges via Lance
1070            // + iterate pending edges in-memory for the `src` column,
1071            // group-by-src. The pending side already includes the row
1072            // we just appended (above).
1073            validate_edge_cardinality_with_pending(self, &ds, staging, &table_key, edge_type)
1074                .await?;
1075
1076            self.invalidate_graph_index().await;
1077
1078            Ok(MutationResult {
1079                affected_nodes: 0,
1080                affected_edges: 1,
1081            })
1082        } else {
1083            Err(OmniError::manifest(format!("unknown type '{}'", type_name)))
1084        }
1085    }
1086
1087    async fn execute_update(
1088        &self,
1089        type_name: &str,
1090        assignments: &[IRAssignment],
1091        predicate: &IRMutationPredicate,
1092        params: &ParamMap,
1093        branch: Option<&str>,
1094        staging: &mut MutationStaging,
1095    ) -> Result<MutationResult> {
1096        // Defense in depth: ensure this is a node type
1097        if !self.catalog().node_types.contains_key(type_name) {
1098            return Err(OmniError::manifest(format!(
1099                "update is only supported for node types, not '{}'",
1100                type_name
1101            )));
1102        }
1103
1104        // Reject updates to @key properties — identity is immutable
1105        if let Some(key_prop) = self.catalog().node_types[type_name].key_property() {
1106            if assignments.iter().any(|a| a.property == key_prop) {
1107                return Err(OmniError::manifest(format!(
1108                    "cannot update @key property '{}' — delete and re-insert instead",
1109                    key_prop
1110                )));
1111            }
1112        }
1113
1114        let pred_sql = predicate_to_sql(predicate, params, false)?;
1115        let schema = self.catalog().node_types[type_name].arrow_schema.clone();
1116        let blob_props = self.catalog().node_types[type_name].blob_properties.clone();
1117
1118        let table_key = format!("node:{}", type_name);
1119        let (ds, _full_path, _table_branch) = open_table_for_mutation(
1120            self,
1121            staging,
1122            branch,
1123            &table_key,
1124            crate::db::MutationOpKind::Update,
1125        )
1126        .await?;
1127
1128        // Scan committed via Lance + apply the same predicate to pending
1129        // batches via DataFusion `MemTable` (read-your-writes for prior
1130        // ops in this query). The pending side may include rows from
1131        // earlier `insert` / `update` ops on the same table.
1132        //
1133        // For blob tables we project away the blob columns: Lance's
1134        // scanner doesn't accept the standard projection path on blob
1135        // descriptors and would panic with a `Field::project` assertion.
1136        // The downstream `apply_assignments` synthesizes blob columns
1137        // from explicit assignments and omits unassigned blobs (Lance's
1138        // merge_insert leaves them untouched). Tables without blob
1139        // columns scan the full schema unprojected.
1140        let non_blob_cols: Vec<&str> = schema
1141            .fields()
1142            .iter()
1143            .filter(|f| !blob_props.contains(f.name()))
1144            .map(|f| f.name().as_str())
1145            .collect();
1146        let projection: Option<&[&str]> =
1147            (!blob_props.is_empty()).then_some(non_blob_cols.as_slice());
1148        let pending_batches = staging.pending_batches(&table_key);
1149        let pending_schema = staging.pending_schema(&table_key);
1150        // Use merge semantics on the union: a committed row whose `id`
1151        // also appears in pending has been logically updated by an
1152        // earlier op in this query and is shadowed from the scan,
1153        // otherwise the predicate runs against stale committed values
1154        // and a chained `update where <pred>` can match a row whose
1155        // pending value no longer satisfies <pred>.
1156        let batches = self
1157            .storage()
1158            .scan_with_pending(
1159                &ds,
1160                pending_batches,
1161                pending_schema,
1162                projection,
1163                Some(&pred_sql),
1164                Some("id"),
1165            )
1166            .await?;
1167
1168        if batches.is_empty() || batches.iter().all(|b| b.num_rows() == 0) {
1169            return Ok(MutationResult {
1170                affected_nodes: 0,
1171                affected_edges: 0,
1172            });
1173        }
1174
1175        // Concat the matched batches (committed + pending) into one. The
1176        // helper trusts that both sides share a schema — Lance returns
1177        // dataset-schema-ordered columns and DataFusion returns
1178        // MemTable-schema-ordered columns; both should match the catalog's
1179        // arrow_schema when the projection is consistent. If they
1180        // diverge (typically a blob-table mid-schema-shift), the helper
1181        // surfaces a clear error directing the caller to split the
1182        // mutation.
1183        let matched = concat_match_batches_to_schema(&schema, &blob_props, batches)?;
1184
1185        let affected_count = matched.num_rows();
1186
1187        let mut resolved: HashMap<String, Literal> = HashMap::new();
1188        for a in assignments {
1189            resolved.insert(a.property.clone(), resolve_expr_value(&a.value, params)?);
1190        }
1191        let updated = apply_assignments(&schema, &matched, &resolved, &blob_props)?;
1192        let node_type = &self.catalog().node_types[type_name];
1193        crate::loader::validate_value_constraints(&updated, node_type)?;
1194        crate::loader::validate_enum_constraints(&updated, &node_type.properties, type_name)?;
1195        let unique_groups = crate::loader::unique_constraint_groups_for_node(node_type);
1196        if !unique_groups.is_empty() {
1197            crate::loader::enforce_unique_constraints_intra_batch(
1198                &updated,
1199                type_name,
1200                &unique_groups,
1201            )?;
1202        }
1203
1204        // Accumulate the updated batch into the Merge-mode pending stream.
1205        // The accumulator may now contain entries with the same id as a
1206        // prior insert or update on this table; `MutationStaging::finalize`
1207        // dedupes by id (last-occurrence wins) before issuing the single
1208        // `stage_merge_insert` call at end-of-query.
1209        let updated_schema = updated.schema();
1210        staging.append_batch(&table_key, updated_schema, PendingMode::Merge, updated)?;
1211
1212        Ok(MutationResult {
1213            affected_nodes: affected_count,
1214            affected_edges: 0,
1215        })
1216    }
1217
1218    async fn execute_delete(
1219        &self,
1220        type_name: &str,
1221        predicate: &IRMutationPredicate,
1222        params: &ParamMap,
1223        branch: Option<&str>,
1224        staging: &mut MutationStaging,
1225    ) -> Result<MutationResult> {
1226        let is_node = self.catalog().node_types.contains_key(type_name);
1227        if is_node {
1228            self.execute_delete_node(type_name, predicate, params, branch, staging)
1229                .await
1230        } else {
1231            self.execute_delete_edge(type_name, predicate, params, branch, staging)
1232                .await
1233        }
1234    }
1235
1236    async fn execute_delete_node(
1237        &self,
1238        type_name: &str,
1239        predicate: &IRMutationPredicate,
1240        params: &ParamMap,
1241        branch: Option<&str>,
1242        staging: &mut MutationStaging,
1243    ) -> Result<MutationResult> {
1244        let pred_sql = predicate_to_sql(predicate, params, false)?;
1245
1246        let table_key = format!("node:{}", type_name);
1247        let (ds, full_path, table_branch) = open_table_for_mutation(
1248            self,
1249            staging,
1250            branch,
1251            &table_key,
1252            crate::db::MutationOpKind::Delete,
1253        )
1254        .await?;
1255        let initial_version = ds.version();
1256
1257        // Scan matching IDs for cascade. Per D₂ this never overlaps with
1258        // staged inserts (mixed insert/delete in one query is rejected at
1259        // parse time), so we scan committed only.
1260        let batches = self
1261            .storage()
1262            .scan(&ds, Some(&["id"]), Some(&pred_sql), None)
1263            .await?;
1264
1265        let deleted_ids: Vec<String> = batches
1266            .iter()
1267            .flat_map(|batch| {
1268                let ids = batch
1269                    .column(0)
1270                    .as_any()
1271                    .downcast_ref::<StringArray>()
1272                    .unwrap();
1273                (0..ids.len())
1274                    .map(|i| ids.value(i).to_string())
1275                    .collect::<Vec<_>>()
1276            })
1277            .collect();
1278
1279        if deleted_ids.is_empty() {
1280            return Ok(MutationResult {
1281                affected_nodes: 0,
1282                affected_edges: 0,
1283            });
1284        }
1285
1286        let affected_nodes = deleted_ids.len();
1287
1288        // Delete nodes — still inline-commit (Lance's `Dataset::delete` is
1289        // not exposed as a two-phase op in 6.0.1). D₂ keeps inserts and
1290        // deletes from coexisting in one query, so this advance of Lance
1291        // HEAD is the only HEAD movement during the query and the
1292        // publisher's CAS captures it intact.
1293        let ds = self
1294            .reopen_for_mutation(
1295                &table_key,
1296                &full_path,
1297                table_branch.as_deref(),
1298                initial_version,
1299                crate::db::MutationOpKind::Delete,
1300            )
1301            .await?;
1302        crate::failpoints::maybe_fail("mutation.delete_node_pre_primary_delete")?;
1303        let (_new_ds, delete_state) = self
1304            .storage_inline_residual()
1305            .delete_where(&full_path, ds, &pred_sql)
1306            .await?;
1307
1308        staging.record_inline(crate::db::SubTableUpdate {
1309            table_key: table_key.clone(),
1310            table_version: delete_state.version,
1311            table_branch: table_branch.clone(),
1312            row_count: delete_state.row_count,
1313            version_metadata: delete_state.version_metadata,
1314        });
1315
1316        let mut affected_edges = 0usize;
1317        let escaped: Vec<String> = deleted_ids
1318            .iter()
1319            .map(|id| format!("'{}'", id.replace('\'', "''")))
1320            .collect();
1321        let id_list = escaped.join(", ");
1322
1323        let edge_info: Vec<(String, String, String)> = self
1324            .catalog()
1325            .edge_types
1326            .iter()
1327            .map(|(name, et)| (name.clone(), et.from_type.clone(), et.to_type.clone()))
1328            .collect();
1329
1330        for (edge_name, from_type, to_type) in &edge_info {
1331            let mut cascade_filters = Vec::new();
1332            if from_type == type_name {
1333                cascade_filters.push(format!("src IN ({})", id_list));
1334            }
1335            if to_type == type_name {
1336                cascade_filters.push(format!("dst IN ({})", id_list));
1337            }
1338            if cascade_filters.is_empty() {
1339                continue;
1340            }
1341
1342            let edge_table_key = format!("edge:{}", edge_name);
1343            let cascade_filter = cascade_filters.join(" OR ");
1344            let (edge_ds, edge_full_path, edge_table_branch) = open_table_for_mutation(
1345                self,
1346                staging,
1347                branch,
1348                &edge_table_key,
1349                crate::db::MutationOpKind::Delete,
1350            )
1351            .await?;
1352
1353            let (_new_edge_ds, edge_delete) = self
1354                .storage_inline_residual()
1355                .delete_where(&edge_full_path, edge_ds, &cascade_filter)
1356                .await?;
1357
1358            affected_edges += edge_delete.deleted_rows;
1359
1360            if edge_delete.deleted_rows > 0 {
1361                staging.record_inline(crate::db::SubTableUpdate {
1362                    table_key: edge_table_key,
1363                    table_version: edge_delete.version,
1364                    table_branch: edge_table_branch,
1365                    row_count: edge_delete.row_count,
1366                    version_metadata: edge_delete.version_metadata,
1367                });
1368            }
1369        }
1370
1371        if affected_edges > 0 {
1372            self.invalidate_graph_index().await;
1373        }
1374
1375        Ok(MutationResult {
1376            affected_nodes,
1377            affected_edges,
1378        })
1379    }
1380
1381    async fn execute_delete_edge(
1382        &self,
1383        type_name: &str,
1384        predicate: &IRMutationPredicate,
1385        params: &ParamMap,
1386        branch: Option<&str>,
1387        staging: &mut MutationStaging,
1388    ) -> Result<MutationResult> {
1389        let pred_sql = predicate_to_sql(predicate, params, true)?;
1390
1391        let table_key = format!("edge:{}", type_name);
1392        let (ds, full_path, table_branch) = open_table_for_mutation(
1393            self,
1394            staging,
1395            branch,
1396            &table_key,
1397            crate::db::MutationOpKind::Delete,
1398        )
1399        .await?;
1400
1401        let (_new_ds, delete_state) = self
1402            .storage_inline_residual()
1403            .delete_where(&full_path, ds, &pred_sql)
1404            .await?;
1405        let affected = delete_state.deleted_rows;
1406
1407        if affected > 0 {
1408            staging.record_inline(crate::db::SubTableUpdate {
1409                table_key,
1410                table_version: delete_state.version,
1411                table_branch,
1412                row_count: delete_state.row_count,
1413                version_metadata: delete_state.version_metadata,
1414            });
1415            self.invalidate_graph_index().await;
1416        }
1417
1418        Ok(MutationResult {
1419            affected_nodes: 0,
1420            affected_edges: affected,
1421        })
1422    }
1423}
1424
1425/// Concat the matched batches from `scan_with_pending` into a single batch.
1426/// `scan_with_pending` returns committed-side and pending-side batches in
1427/// order; both should share a schema if pending was produced through
1428/// `apply_assignments` with full-schema scan input. If schemas drift,
1429/// surface a clear error so the user can split the query.
1430fn concat_match_batches_to_schema(
1431    _schema: &SchemaRef,
1432    _blob_properties: &HashSet<String>,
1433    batches: Vec<RecordBatch>,
1434) -> Result<RecordBatch> {
1435    if batches.len() == 1 {
1436        return Ok(batches.into_iter().next().unwrap());
1437    }
1438    let common = batches[0].schema();
1439    arrow_select::concat::concat_batches(&common, &batches).map_err(|e| {
1440        OmniError::Lance(format!(
1441            "scan_with_pending returned batches with mismatched schemas \
1442             across the committed/pending boundary; this typically indicates \
1443             a blob-column shape mismatch between the committed table and a \
1444             prior in-query insert/update. Split blob-touching mutations \
1445             into separate queries. ({})",
1446            e
1447        ))
1448    })
1449}
1450
1451/// Validate `@card` bounds against committed (Lance) + pending (in-memory)
1452/// edges for one edge table. Engine path: each insert produces a fresh
1453/// ULID id, so committed and pending cannot share a primary key — no
1454/// dedup needed (`dedupe_key_column = None`).
1455async fn validate_edge_cardinality_with_pending(
1456    db: &Omnigraph,
1457    committed_ds: &SnapshotHandle,
1458    staging: &MutationStaging,
1459    table_key: &str,
1460    edge_type: &omnigraph_compiler::catalog::EdgeType,
1461) -> Result<()> {
1462    if edge_type.cardinality.is_default() {
1463        return Ok(());
1464    }
1465    let counts =
1466        super::staging::count_src_per_edge(db, committed_ds, table_key, staging, None).await?;
1467    super::staging::enforce_cardinality_bounds(edge_type, &counts)
1468}
1469
1470fn enrich_mutation_params(params: &ParamMap) -> Result<ParamMap> {
1471    let mut resolved = params.clone();
1472    if !resolved.contains_key(NOW_PARAM_NAME) {
1473        let now = OffsetDateTime::now_utc()
1474            .format(&Rfc3339)
1475            .map_err(|e| OmniError::manifest(format!("failed to format now(): {}", e)))?;
1476        resolved.insert(NOW_PARAM_NAME.to_string(), Literal::DateTime(now));
1477    }
1478    Ok(resolved)
1479}