#excel #xlsx #batch #reader #low-memory #file-reader #read-file

xlsx_batch_reader

An Excel file(xlsx/xlsm) reader by batches, in pure Rust

14 releases

new 0.2.0 Nov 5, 2024
0.1.14 Nov 3, 2024
0.1.13 Sep 17, 2024
0.1.12 Jun 1, 2024
0.1.7 Apr 27, 2024

#1006 in Parser implementations

Download history 102/week @ 2024-07-30 1/week @ 2024-09-10 159/week @ 2024-09-17 29/week @ 2024-09-24 15/week @ 2024-10-01 76/week @ 2024-10-29

76 downloads per month

MIT license

94KB
1.5K SLoC

XlsxBatchReader

An Excel/OpenDocument Spreadsheets file batch reader, in pure Rust. This crate supports Office 2007 or newer file formats(xlsx, xlsm, etc). The most obvious difference from other Excel file reading crates is that it does not read the whole file into memory, but read in batches. So that it can maintain low memory usage, especially when reading large files. This crate supports date and time recognition, as well as obtaining merged cell ranges. For faster speed, it only supports reading data, not support formulas and other styles.

Examples

  1. simple reader
use xlsx_batch_reader::{get_ord_from_tuple, read::XlsxBook, MAX_COL_NUM};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // the tail empty rows will be ignored, if you want the length of cells in each row is fixed, you can set right_ncol to a number not MAX_COL_NUM
        let sheet = book.get_sheet_by_name(&shname, 100, 0, 1, MAX_COL_NUM, false)?;

        for batch in sheet {
            let (rows_nums, rows_data) = batch?;
            // empty rows will be skiped
            for (row, cells) in rows_nums.into_iter().zip(rows_data) {
                for (col, cel) in cells.into_iter().enumerate() {
                    // supprted types: String, i64, f64, bool, Date32, Timestamp(v0.1.4), NaiveDate, NaiveDateTime(v0.1.2), NaiveTime(v0.1.2)
                    let val: String = cel.get()?.unwrap();   
                    println!("the value of {} is {val}; raw cell is {:?}", get_ord_from_tuple(row, (col+1) as u16)?, cel);   
                }
            }
        };
    }
    Ok(())
}

possible output:

the value of A1 is a; raw cell is Shared("a")
the value of B1 is ; raw cell is Blank
the value of C1 is c; raw cell is Shared("c")
the value of D1 is d; raw cell is Shared("d")
the value of A2 is 1; raw cell is Number(1.0)
the value of B2 is ; raw cell is Blank
the value of C2 is s; raw cell is Shared("s")
the value of A4 is 2024-01-04; raw cell is Date(45295.58405092593)
the value of B4 is ; raw cell is Blank
the value of C4 is 4; raw cell is Number(4.0)  
  1. merged ranges
use xlsx_batch_reader::{get_num_from_ord, is_merged_cell, read::XlsxBook};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // each row will have 3 cells.
        let mut sheet = book.get_sheet_by_name(&shname, 100, 0, 1, get_num_from_ord("C".as_bytes())?, true)?;

        // this is not necessary, if you don't care about the headers.
        let (_, _header) = sheet.get_header_row()?;
        if let Some((_rows_nums, _rows_data)) = sheet.get_remaining_cells()? {
            //  some code
        }; 

        // should be called when all data have been scaned.
        let merged_rngs = sheet.get_merged_ranges()?;
        match is_merged_cell(merged_rngs, 2, get_num_from_ord("A".as_bytes())?) {
            (true, None) => {
                println!("a merged cell(not top left cell)");
            },
            (true, Some((nrow, ncol))) => {
                println!("a merged cell(top left cell), taking {nrow} row(s) and {ncol} column(s)");
            },
            _ => {
                println!("not a merged cell");
            }
        }
    }
    Ok(())
}

possible output:

a merged cell(top left cell), taking 2 row(s) and 2 column(s)
  1. read date and time
use chrono::{NaiveDate, NaiveDateTime, NaiveTime};
use xlsx_batch_reader::{read::XlsxBook, Timestamp, MAX_COL_NUM};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // the tail empty cells will be ignored, if you want the length of cells in each row is fixed, you can set right_ncol to a number not MAX_COL_NUM
        let mut sheet = book.get_sheet_by_name(&shname, 100, 3, 1, MAX_COL_NUM, false)?;

        if let Some((_, rows_data)) = sheet.get_remaining_cells()? {
            let row = &rows_data[0];
            let val_dt: NaiveDate = row[0].get()?.unwrap();
            let val_tm: NaiveTime = row[0].get()?.unwrap();
            let val_dttm: NaiveDateTime = row[0].get()?.unwrap();
            let val_stamp: Timestamp = row[0].get()?.unwrap();   // since v0.1.4
            println!("date:{}\ntime:{}\ndatetime:{}\ntimestamp:{}", val_dt, val_tm, val_dttm, val_stamp.utc());
        }; 
    }
    Ok(())
}

possible output:

date:2024-01-04
time:14:01:02
datetime:2024-01-04 14:01:02
timestamp:1704376862
  1. read partial rows based on conditions
use std::collections::HashMap;
use xlsx_batch_reader::{get_ord_from_tuple, read::XlsxBook, MAX_COL_NUM};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // the tail empty cells will be ignored, if you want the length of cells in each row is fixed, you can set right_ncol to a number not MAX_COL_NUM
        let mut sheet = book.get_sheet_by_name(&shname, 100, 0, 1, MAX_COL_NUM, true)?;

        let mut skip_until = HashMap::new();
        skip_until.insert("A".into(), "col1".into());
        skip_until.insert("C".into(), "col3".into());
        sheet.with_skip_until(&skip_until);
        let mut read_before = HashMap::new();
        read_before.insert("B".into(), "sum".into());
        sheet.with_read_before(&read_before);
        // only the rows after skip_until(included) and before read_before(excluded) will be returned

        for batch in sheet {
            let (rows_nums, rows_data) = batch?;
            // empty rows will be skiped
            for (row, cells) in rows_nums.into_iter().zip(rows_data) {
                for (col, cel) in cells.into_iter().enumerate() {
                    let val: String = cel.get()?.unwrap();   
                    println!("the value of {} is {val}; raw cell is {:?}", get_ord_from_tuple(row, (col+1) as u16)?, cel);  
                }
            }
        };
    }
  1. cached reader(feature cached should be enabled)
use xlsx_batch_reader::{read::XlsxBook, MAX_COL_NUM};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // iter_batch will be supported in the future
        // the tail empty cells will be ignored, if you want the length of cells in each row is fixed, you can set right_ncol to a number not MAX_COL_NUM
        let sheet = book.get_cached_sheet_by_name(&shname, 100, 1, 1, MAX_COL_NUM, false)?;

        println!("sheet: {}, row_ranges: {:?}, col_ranges: {:?}", sheet.sheet_name(), sheet.row_range(), sheet.column_range());

        let (_, merge_info) = sheet.get_cell_value_with_merge_info("B2")?;

        match merge_info {
            (true, None) => {
                println!("B2 is a merged cell(not top left cell)");
            },
            (true, Some((nrow, ncol))) => {
                println!("B2 is a merged cell(top left cell), taking {nrow} row(s) and {ncol} column(s)");
            },
            _ => {
                println!("B2 is not a merged cell");
            }
        };

        let a4 = sheet.get_cell_value("A4")?;
        println!("A4={:?}", a4);
    }
    Ok(())
}

possible output:

sheet: Sheet1, row_ranges: (2, 4), col_ranges: (1, 4)
B2 is a merged cell(not top left cell)
A4=Date(45295.58405092593)
  1. simple batch writer (feature xlsxwriter should be enabled)
use xlsx_batch_reader::{get_num_from_ord, read::XlsxBook, write::XlsxWriter};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut writer = XlsxWriter::new();
    let mut book = XlsxBook::new("xlsx/test.xlsx", true)?;
    for shname in book.get_visible_sheets().clone() {
        // left_ncol should not be 0
        // each row will have 3 cells.
        let mut sheet = book.get_sheet_by_name(&shname, 100, 0, 1, get_num_from_ord("C".as_bytes())?, true)?;

        // the sheet name will be write at the begin of each row
        let pre_cells = vec![shname];
        if let Some((rows_nums, rows_data)) = sheet.get_remaining_cells()? {
            writer.append_rows("sheet", rows_nums, rows_data, &pre_cells)?;
            // if you don't want row numbers to be writed before data, set nrows = vec![];
        }; 
    };
    writer.save_as("xlsx/out.xlsx")?;
    Ok(())
}
  1. write row(s) by column name (feature xlsxwriter should be enabled)
use std::collections::HashMap;
use xlsx_batch_reader::write::XlsxWriter;

fn main() -> Result<(), Box<dyn std::error::Error>> {

    let mut writer = XlsxWriter::new();
    // if you have many sheets call this for each sheet
    writer.with_columns("Sheet1".to_string(), vec!["A".to_string(), "B".to_string(), "C".to_string(), "D".to_string()], true);

    let row: HashMap<String, i32> = vec![("A".to_string(), 1), ("C".to_string(), 3)].into_iter().collect();
    writer.append_row_by_name("Sheet1", row)?;
    let row1: HashMap<String, &str> = vec![("A".to_string(), "A3"), ("B".to_string(), "B3"), ("D".to_string(), "D3")].into_iter().collect();
    let row2: HashMap<String, &str> = vec![("B".to_string(), "B4"), ("C".to_string(), "C4")].into_iter().collect();
    writer.append_rows_by_name("Sheet1", vec![row1, row2])?;

    writer.save_as("xlsx/out.xlsx")?;
    Ok(())
}

Features

Features Description
cached Enable cached xlsx sheet (read all sheet data into memory when created)
xlsxwriter Enable simple xlsx writer(depends on rust_xlsxwriter crate)
full Enable feature cached and xlsxwriter

Todos

  • get cell with extra info

Changelogs

please refer to 'changelogs.md'

Dependencies

~7–14MB
~150K SLoC