excelerate 0.14.0

Read, write and recalculate spreadsheets: xlsx, xls, ods, csv, html, and a 518-function formula engine
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
//! Summaries of a table in one formula: `GROUPBY` and `PIVOTBY`, with two
//! small relatives, `PERCENTOF` and `TRIMRANGE`.
//!
//! The summarising function is a lambda or a bare function name, as Excel
//! lets it be written: `GROUPBY(A2:A9, C2:C9, SUM)`. A bare name is not a value
//! anywhere else in this engine, so these two look for it themselves.
//!
//! The layout - headers, where the totals go, how groups sort - follows
//! Microsoft's description of the functions. No other engine at hand
//! implements them, so it has not been checked against Excel itself.

use super::{Arg, aggregate_numbers, first_error, is_known};
use crate::error::CellError;
use crate::formula::eval::{Engine, Origin};
use crate::formula::parser::Expr;
use crate::formula::value::{Value, compare};
use std::cmp::Ordering;

/// What a group of values is summarised with.
enum Summary {
    Lambda(Value),
    Builtin(String),
}

/// The name a bare function's argument is bound to while it runs.
const ETA: &str = "_xleta.values";

impl Summary {
    fn read(engine: &mut Engine<'_>, origin: Origin, expr: &Expr) -> Result<Self, Value> {
        let value = engine.eval_expr(origin, expr);
        match (&value, expr) {
            (Value::Lambda(_), _) => Ok(Self::Lambda(value)),
            (Value::Error(CellError::Name), Expr::Name(name)) if is_known(&bare(name)) => {
                Ok(Self::Builtin(bare(name)))
            }
            (Value::Error(e), _) => Err(Value::Error(*e)),
            _ => Err(Value::Error(CellError::Calc)),
        }
    }

    fn apply(&self, engine: &mut Engine<'_>, origin: Origin, values: Vec<Value>) -> Value {
        let column = Value::array(values.into_iter().map(|v| vec![v]).collect());
        let out = match self {
            Self::Lambda(f) => engine.apply(origin, f, vec![column]),
            Self::Builtin(name) => engine.scoped(vec![(ETA.to_owned(), column)], |engine| {
                super::call(engine, origin, name, &[Expr::Name(ETA.to_owned())])
            }),
        };
        out.scalar().clone()
    }
}

/// A function name written as a value, without the prefix a file stores it
/// under: Excel saves `GROUPBY(A2:A9,C2:C9,SUM)` as `..._xleta.SUM)`.
fn bare(name: &str) -> String {
    let upper = name.to_ascii_uppercase();
    let mut rest = upper.as_str();
    for prefix in ["_XLFN.", "_XLETA."] {
        rest = rest.strip_prefix(prefix).unwrap_or(rest);
    }
    rest.to_owned()
}

/// A value as rows of cells.
fn grid(value: Value) -> Vec<Vec<Value>> {
    match value {
        Value::Array(rows) => std::sync::Arc::unwrap_or_clone(rows),
        other => vec![vec![other]],
    }
}

/// An optional numeric argument, `default` when left out.
fn option(
    engine: &mut Engine<'_>,
    origin: Origin,
    args: &[Expr],
    at: usize,
    default: f64,
) -> Result<f64, Value> {
    match args.get(at) {
        None | Some(Expr::Missing) => Ok(default),
        Some(expr) => engine
            .eval_expr(origin, expr)
            .scalar()
            .number()
            .map(f64::trunc)
            .map_err(Value::Error),
    }
}

/// Header handling: whether the first row is headers, and whether to show them.
#[derive(Clone, Copy)]
struct Headers {
    present: bool,
    shown: bool,
    generated: bool,
}

impl Headers {
    /// Reads `field_headers`. Left out, the first row counts as headers when
    /// the first value is text and the one under it a number, and they are
    /// not shown.
    fn read(mode: Option<f64>, values: &[Vec<Value>]) -> Result<Self, Value> {
        let mode = mode.unwrap_or_else(|| {
            let first = values.first().and_then(|r| r.first());
            let second = values.get(1).and_then(|r| r.first());
            let looks =
                matches!(first, Some(Value::Text(_))) && matches!(second, Some(Value::Number(_)));
            if looks { 1.0 } else { 0.0 }
        });
        Ok(match mode {
            0.0 => Self {
                present: false,
                shown: false,
                generated: false,
            },
            1.0 => Self {
                present: true,
                shown: false,
                generated: false,
            },
            2.0 => Self {
                present: false,
                shown: true,
                generated: true,
            },
            3.0 => Self {
                present: true,
                shown: true,
                generated: false,
            },
            _ => return Err(Value::Error(CellError::Value)),
        })
    }

    /// The header row taken off `rows`, or names made up for its columns.
    fn take(self, rows: &mut Vec<Vec<Value>>, width: usize, stem: &str) -> Vec<Value> {
        if self.present && !rows.is_empty() {
            return rows.remove(0);
        }
        (1..=width)
            .map(|i| {
                Value::Text(if self.generated {
                    format!("{stem} {i}")
                } else {
                    String::new()
                })
            })
            .collect()
    }
}

/// The rows the filter keeps, all of them when there is none.
fn kept(
    engine: &mut Engine<'_>,
    origin: Origin,
    filter: Option<&Expr>,
    height: usize,
    headers: bool,
) -> Result<Vec<bool>, Value> {
    let Some(expr) = filter.filter(|e| !matches!(e, Expr::Missing)) else {
        return Ok(vec![true; height]);
    };
    let mut flat: Vec<Value> = grid(engine.eval_expr(origin, expr))
        .into_iter()
        .flatten()
        .collect();
    // A filter may cover the header row or not.
    if headers && flat.len() == height + 1 {
        flat.remove(0);
    }
    if flat.len() != height {
        return Err(Value::Error(CellError::Value));
    }
    flat.iter()
        .map(|v| match v {
            Value::Error(e) => Err(Value::Error(*e)),
            other => Ok(other.boolean().unwrap_or(false)),
        })
        .collect()
}

/// Keys equal the way Excel groups them: by value, text without case.
fn same_key(a: &[Value], b: &[Value]) -> bool {
    a.len() == b.len()
        && a.iter()
            .zip(b)
            .all(|(x, y)| compare(x, y) == Ordering::Equal)
}

/// The distinct keys, in order of first appearance.
fn distinct(keys: &[Vec<Value>], rows: &[usize]) -> Vec<Vec<Value>> {
    let mut out: Vec<Vec<Value>> = Vec::new();
    for &r in rows {
        if !out.iter().any(|k| same_key(k, &keys[r])) {
            out.push(keys[r].clone());
        }
    }
    out
}

fn total_label(width: usize) -> Vec<Value> {
    let mut row = vec![Value::Blank; width];
    if let Some(first) = row.first_mut() {
        *first = Value::Text("Total".into());
    }
    row
}

/// `GROUPBY(row_fields, values, function, [field_headers], [total_depth],
/// [sort_order], [filter_array], [field_relationship])`
pub fn groupby(engine: &mut Engine<'_>, origin: Origin, args: &[Expr]) -> Value {
    match groupby_inner(engine, origin, args) {
        Ok(v) | Err(v) => v,
    }
}

fn groupby_inner(engine: &mut Engine<'_>, origin: Origin, args: &[Expr]) -> Result<Value, Value> {
    if args.len() < 3 || args.len() > 8 {
        return Err(Value::Error(CellError::Value));
    }
    let mut fields = grid(engine.eval_expr(origin, &args[0]));
    let mut values = grid(engine.eval_expr(origin, &args[1]));
    let summary = Summary::read(engine, origin, &args[2])?;
    if let Some(e) = error_in(&fields).or_else(|| error_in(&values)) {
        return Err(Value::Error(e));
    }
    if fields.len() != values.len() || fields.is_empty() {
        return Err(Value::Error(CellError::Value));
    }
    let mode = match args.get(3) {
        None | Some(Expr::Missing) => None,
        Some(_) => Some(option(engine, origin, args, 3, 0.0)?),
    };
    let headers = Headers::read(mode, &values)?;
    let depth = option(engine, origin, args, 4, 1.0)?;
    let sort = option(engine, origin, args, 5, 1.0)?;
    let width = fields[0].len();
    let value_width = values[0].len();
    let field_names = headers.take(&mut fields, width, "Row Field");
    let value_names = headers.take(&mut values, value_width, "Value");
    let height = fields.len();
    let keep = kept(engine, origin, args.get(6), height, headers.present)?;
    let rows: Vec<usize> = (0..height).filter(|&r| keep[r]).collect();

    let mut ctx = Grouping {
        engine,
        origin,
        summary: &summary,
        fields: &fields,
        values: &values,
        depth,
        sort,
    };
    let mut out = Vec::new();
    if headers.shown {
        out.push(field_names.into_iter().chain(value_names).collect());
    }
    let grand = |ctx: &mut Grouping<'_, '_>| {
        let mut row = total_label(width);
        row.extend(ctx.summarise(&rows));
        row
    };
    if depth < 0.0 {
        out.push(grand(&mut ctx));
    }
    ctx.level(0, &rows, &mut out);
    if depth > 0.0 {
        out.push(grand(&mut ctx));
    }
    Ok(Value::array(out))
}

fn error_in(rows: &[Vec<Value>]) -> Option<CellError> {
    rows.iter().flatten().find_map(|v| match v {
        Value::Error(e) => Some(*e),
        _ => None,
    })
}

/// What one `GROUPBY` works with while it lays the groups out.
struct Grouping<'e, 'a> {
    engine: &'e mut Engine<'a>,
    origin: Origin,
    summary: &'e Summary,
    fields: &'e [Vec<Value>],
    values: &'e [Vec<Value>],
    depth: f64,
    sort: f64,
}

impl Grouping<'_, '_> {
    /// The summary of each value column over `rows`.
    fn summarise(&mut self, rows: &[usize]) -> Vec<Value> {
        let width = self.values.first().map_or(0, Vec::len);
        (0..width)
            .map(|c| {
                let column = rows.iter().map(|&r| self.values[r][c].clone()).collect();
                self.summary.apply(self.engine, self.origin, column)
            })
            .collect()
    }

    /// Lays out the groups of field `level` among `rows`.
    fn level(&mut self, level: usize, rows: &[usize], out: &mut Vec<Vec<Value>>) {
        let width = self.fields.first().map_or(0, Vec::len);
        let keys: Vec<Vec<Value>> = self.fields.iter().map(|r| vec![r[level].clone()]).collect();
        let mut groups: Vec<(Vec<Value>, Vec<usize>, Vec<Value>)> = distinct(&keys, rows)
            .into_iter()
            .map(|key| {
                let members: Vec<usize> = rows
                    .iter()
                    .copied()
                    .filter(|&r| same_key(&keys[r], &key))
                    .collect();
                let summary = self.summarise(&members);
                (key, members, summary)
            })
            .collect();
        // The sort column counts the fields and then the values, from one.
        #[expect(
            clippy::cast_possible_truncation,
            clippy::cast_sign_loss,
            reason = "a small column index"
        )]
        let column = self.sort.abs() as usize;
        let descending = self.sort < 0.0;
        groups.sort_by(|a, b| {
            let ordering = if column > width {
                let c = column - width - 1;
                match (a.2.get(c), b.2.get(c)) {
                    (Some(x), Some(y)) => compare(x, y),
                    _ => Ordering::Equal,
                }
            } else {
                compare(&a.0[0], &b.0[0])
            };
            if descending && (column == level + 1 || column > width) {
                ordering.reverse()
            } else {
                ordering
            }
        });
        let last = level + 1 == width;
        #[expect(
            clippy::cast_possible_truncation,
            clippy::cast_sign_loss,
            reason = "a small depth"
        )]
        let depth = self.depth.abs() as usize;
        let subtotal = !last && level + 1 < depth;
        for (_, members, summary) in groups {
            let prefix = |fields: &[Vec<Value>]| -> Vec<Value> {
                let first = members[0];
                (0..width)
                    .map(|c| {
                        if c <= level {
                            fields[first][c].clone()
                        } else {
                            Value::Blank
                        }
                    })
                    .collect()
            };
            let total_row = || {
                let mut row = prefix(self.fields);
                row.extend(summary.iter().cloned());
                row
            };
            if subtotal && self.depth < 0.0 {
                out.push(total_row());
            }
            if last {
                out.push(total_row());
            } else {
                self.level(level + 1, &members, out);
            }
            if subtotal && self.depth > 0.0 {
                let mut row = prefix(self.fields);
                row.extend(summary.iter().cloned());
                out.push(row);
            }
        }
    }
}

/// `PIVOTBY(row_fields, col_fields, values, function, [field_headers],
/// [row_total_depth], [row_sort_order], [col_total_depth], [col_sort_order],
/// [filter_array], [relative_to])`
///
/// ponytail: totals are grand totals only; a depth of two or more, which asks
/// for subtotals, is read as one.
pub fn pivotby(engine: &mut Engine<'_>, origin: Origin, args: &[Expr]) -> Value {
    match pivotby_inner(engine, origin, args) {
        Ok(v) | Err(v) => v,
    }
}

fn pivotby_inner(engine: &mut Engine<'_>, origin: Origin, args: &[Expr]) -> Result<Value, Value> {
    if args.len() < 4 || args.len() > 11 {
        return Err(Value::Error(CellError::Value));
    }
    let mut row_fields = grid(engine.eval_expr(origin, &args[0]));
    let mut col_fields = grid(engine.eval_expr(origin, &args[1]));
    let mut values = grid(engine.eval_expr(origin, &args[2]));
    let summary = Summary::read(engine, origin, &args[3])?;
    for g in [&row_fields, &col_fields, &values] {
        if let Some(e) = error_in(g) {
            return Err(Value::Error(e));
        }
    }
    if row_fields.len() != values.len() || col_fields.len() != values.len() || values.is_empty() {
        return Err(Value::Error(CellError::Value));
    }
    let mode = match args.get(4) {
        None | Some(Expr::Missing) => None,
        Some(_) => Some(option(engine, origin, args, 4, 0.0)?),
    };
    let headers = Headers::read(mode, &values)?;
    let row_depth = option(engine, origin, args, 5, 1.0)?;
    let row_sort = option(engine, origin, args, 6, 1.0)?;
    let col_depth = option(engine, origin, args, 7, 1.0)?;
    let col_sort = option(engine, origin, args, 8, 1.0)?;
    let (row_width, col_width, value_width) =
        (row_fields[0].len(), col_fields[0].len(), values[0].len());
    let row_names = headers.take(&mut row_fields, row_width, "Row Field");
    let _ = headers.take(&mut col_fields, col_width, "Col Field");
    let value_names = headers.take(&mut values, value_width, "Value");
    let height = values.len();
    let keep = kept(engine, origin, args.get(9), height, headers.present)?;
    let rows: Vec<usize> = (0..height).filter(|&r| keep[r]).collect();

    let row_keys = sorted_keys(&row_fields, &rows, row_sort);
    let col_keys = sorted_keys(&col_fields, &rows, col_sort);

    // Each column of the report: a column key (or the total) and a value.
    let mut columns: Vec<(Option<&Vec<Value>>, usize)> = Vec::new();
    let totals_column: Vec<(Option<&Vec<Value>>, usize)> = if col_depth == 0.0 {
        Vec::new()
    } else {
        (0..value_width).map(|v| (None, v)).collect()
    };
    if col_depth < 0.0 {
        columns.extend(totals_column.iter().copied());
    }
    for key in &col_keys {
        columns.extend((0..value_width).map(|v| (Some(key), v)));
    }
    if col_depth > 0.0 {
        columns.extend(totals_column.iter().copied());
    }

    let names = headers.shown.then_some((row_names, value_names.as_slice()));
    let mut out = column_headers(&columns, row_width, col_width, value_width, names);

    let mut cell = |row_key: Option<&Vec<Value>>, col_key: Option<&Vec<Value>>, value: usize| {
        let members: Vec<Value> = rows
            .iter()
            .copied()
            .filter(|&r| row_key.is_none_or(|k| same_key(k, &row_fields[r])))
            .filter(|&r| col_key.is_none_or(|k| same_key(k, &col_fields[r])))
            .map(|r| values[r][value].clone())
            .collect();
        if members.is_empty() {
            Value::Blank
        } else {
            summary.apply(engine, origin, members)
        }
    };
    let mut body: Vec<Vec<Value>> = Vec::new();
    for key in &row_keys {
        let mut row = key.clone();
        for (col, value) in &columns {
            row.push(cell(Some(key), *col, *value));
        }
        body.push(row);
    }
    let mut total = total_label(row_width);
    for (col, value) in &columns {
        total.push(cell(None, *col, *value));
    }
    if row_depth < 0.0 {
        out.push(total);
        out.extend(body);
    } else {
        out.extend(body);
        if row_depth > 0.0 {
            out.push(total);
        }
    }
    Ok(Value::array(out))
}

/// The distinct keys among `rows`, sorted by the one-based key column
/// `order` names, descending when it is negative.
fn sorted_keys(keys: &[Vec<Value>], rows: &[usize], order: f64) -> Vec<Vec<Value>> {
    let mut distinct = distinct(keys, rows);
    #[expect(
        clippy::cast_possible_truncation,
        clippy::cast_sign_loss,
        reason = "a small column index"
    )]
    let column = (order.abs() as usize).saturating_sub(1);
    distinct.sort_by(|a, b| {
        let o = match (a.get(column), b.get(column)) {
            (Some(x), Some(y)) => compare(x, y),
            _ => Ordering::Equal,
        };
        if order < 0.0 { o.reverse() } else { o }
    });
    distinct
}

/// The header rows of a `PIVOTBY`: a row per column field with its keys over
/// the columns, and a row of value names when there are several values or the
/// headers are shown (`names`: the row field names and the value names).
fn column_headers(
    columns: &[(Option<&Vec<Value>>, usize)],
    row_width: usize,
    col_width: usize,
    value_width: usize,
    names: Option<(Vec<Value>, &[Value])>,
) -> Vec<Vec<Value>> {
    let mut out: Vec<Vec<Value>> = Vec::new();
    for level in 0..col_width {
        let mut row = vec![Value::Blank; row_width];
        for (key, value) in columns {
            row.push(match key {
                Some(k) if *value == 0 => k[level].clone(),
                None if *value == 0 && level == 0 => Value::Text("Total".into()),
                _ => Value::Blank,
            });
        }
        out.push(row);
    }
    if value_width > 1 || names.is_some() {
        let (mut row, value_names) = match names {
            Some((rows, values)) => (rows, Some(values)),
            None => (vec![Value::Blank; row_width], None),
        };
        row.extend(
            columns
                .iter()
                .map(|(_, v)| value_names.map_or(Value::Blank, |n| n[*v].clone())),
        );
        out.push(row);
    }
    out
}

/// `PERCENTOF(data_subset, data_all)`
pub fn percentof(args: &[Arg]) -> Value {
    let [subset, all] = args else {
        return Value::Error(CellError::Value);
    };
    if let Some(e) = first_error(args) {
        return Value::Error(e);
    }
    let (Ok(part), Ok(whole)) = (
        aggregate_numbers(std::slice::from_ref(subset)),
        aggregate_numbers(std::slice::from_ref(all)),
    ) else {
        return Value::Error(CellError::Value);
    };
    let whole: f64 = whole.iter().sum();
    if whole == 0.0 {
        return Value::Error(CellError::Div0);
    }
    Value::Number(part.iter().sum::<f64>() / whole)
}

/// `TRIMRANGE(range, [trim_rows], [trim_cols])`: the range without its empty
/// outer rows and columns. 0 trims nothing, 1 the leading ones, 2 the
/// trailing ones, 3 (the default) both.
pub fn trimrange(args: &[Arg]) -> Value {
    let Some((range, rest)) = args.split_first() else {
        return Value::Error(CellError::Value);
    };
    if rest.len() > 2 {
        return Value::Error(CellError::Value);
    }
    let how = |i: usize| -> Result<u8, CellError> {
        match rest.get(i) {
            None => Ok(3),
            Some(a) if a.missing() => Ok(3),
            Some(a) => match a.number()?.trunc() {
                #[expect(
                    clippy::cast_possible_truncation,
                    clippy::cast_sign_loss,
                    reason = "checked to be 0..=3"
                )]
                n if (0.0..=3.0).contains(&n) => Ok(n as u8),
                _ => Err(CellError::Value),
            },
        }
    };
    let (rows_how, cols_how) = match (how(0), how(1)) {
        (Ok(r), Ok(c)) => (r, c),
        (Err(e), _) | (_, Err(e)) => return Value::Error(e),
    };
    let grid = grid(range.value.clone());
    let blank = |v: &Value| matches!(v, Value::Blank);
    let height = grid.len();
    let width = grid.first().map_or(0, Vec::len);
    let row_empty = |r: usize| grid[r].iter().all(blank);
    let col_empty = |c: usize| grid.iter().all(|row| row.get(c).is_none_or(blank));
    let span = |len: usize, how: u8, empty: &dyn Fn(usize) -> bool| {
        let mut start = 0;
        let mut end = len;
        if how & 1 == 1 {
            while start < end && empty(start) {
                start += 1;
            }
        }
        if how & 2 == 2 {
            while end > start && empty(end - 1) {
                end -= 1;
            }
        }
        start..end
    };
    let rows = span(height, rows_how, &row_empty);
    let cols = span(width, cols_how, &col_empty);
    if rows.is_empty() || cols.is_empty() {
        return Value::Error(CellError::Calc);
    }
    Value::array(
        grid[rows]
            .iter()
            .map(|row| row[cols.clone()].to_vec())
            .collect(),
    )
}