Crate cindex

source ·
Expand description

Cindex, a csv indexer

Cindex is a easy to use csv indexer with SQL-like simple query support.

Cindex is not intended for heavy database indexing but for simple in-memory querying. Use other databases interaction layer if you’re using big chunks of csv files.

Usage

[dependencies]
cindex = "*" # Use the latest version if possible

# Use "default-features = false" if you don't need rayon iteration enhancement.

use std::fs::File;
use std::io::BufReader;
use std::str::FromStr;
use cindex::{Indexer, Predicate, Query, OutOption, Operator};

let mut indexer = Indexer::new();

// Add table without types
// Default types for every columns are "Text"
indexer
    .add_table(
        "table1",
        BufReader::new(File::open("test.csv").expect("Failed to open a file")),
    )
    .expect("Failed to add table");

// Add table from stdin
let stdin = std::io::stdin();
indexer
    .add_table("table3", stdin.lock())
    .expect("Failed to add table");

// Indexing

// Create query object and print output to terminal
let query = Query::from_str("SELECT a,b,c FROM table1 WHERE a = 10").expect("Failed to create query");
indexer
    .index(query, OutOption::Term)
    .expect("Failed to index a table");

// Use raw query and yield output to a file
indexer
    .index_raw(
        "SELECT * FROM table3 WHERE id = 10",
        OutOption::File(std::fs::File::create("out.csv").expect("Failed to create a file")),
    )
    .expect("Failed to index a table");

// Use builder pattern to construct query and index a table
let query = Query::build()
    .table("table2")
    .columns(vec!["id", "address"])
    .predicate(Predicate::new("id", Operator::Equal).args(vec!["10"]))
    .predicate(
        Predicate::build()
            .column("address")
            .operator(Operator::NotEqual)
            .raw_args("111-2222"),
    );

let mut acc = String::new();
indexer
    .index(query, OutOption::Value(&mut acc))
    .expect("Failed to index a table");

// Always use unix newline for formatting
indexer.always_use_unix_newline(true);

Query

Cindex’s query syntax is similar to SQL but has some small differences.

/* Select everythig from given table*/
SELECT * FROM table1

/* Select everything from given table and order by column with descending
order*/
SELECT * FROM table1 ORDER BY col1 DESC

/* Same with previous commands but map headers into custom values */
SELECT * FROM table1 ORDER BY col1 DESC HMAP new_h,new_h2,new_h3

/* -> Previous query result header with underbar substituted with whitespaces
new h,new h2,new h3
<-- Content goes here -->
 */

/* Select given columns from table where column's value is equal to given
condition and also other column's value matches regex expression */
SELECT col1,col2 FROM table1 WHERE col1 = 10 AND col2 LIKE ^start

/* There is a flag syntax which changes query behaviour*/
SELECT * FROM table_name FLAG PHD SUP

/* In this case each flag does next operation
  - PHD (PRINT-HEADER) : Print header in result output
  - SUP (SUPPLEMENT)   : Enable selection of non-existent column with empty values
  - TP  (Tranpose)     : Tranpose result
 */

Supported WHERE operations are

 >=
 >
 <=
 <
 =
 !=
 IN ( enumerate )
 BETWEEN (inclusive range of min & max)
 LIKE ( with regeular expression )

Structs

  • Indexer
    Entry struct for indexing csv tables
  • Predicate
    Predicate to decide whether a specific row qualifies a query or not
  • Query
    Query to index a table
  • ReaderOption
    Reader behaviour related options

Enums

Type Aliases