70 releases
0.26.1 | Oct 10, 2024 |
---|---|
0.25.0 | May 24, 2024 |
0.24.0 | Feb 8, 2024 |
0.23.1 | Dec 19, 2023 |
0.3.2 | Nov 27, 2016 |
#60 in Parser implementations
103,920 downloads per month
Used in 82 crates
(57 directly)
375KB
9K
SLoC
calamine
An Excel/OpenDocument Spreadsheets file reader/deserializer, in pure Rust.
Description
calamine is a pure Rust library to read and deserialize any spreadsheet file:
- excel like (
xls
,xlsx
,xlsm
,xlsb
,xla
,xlam
) - opendocument spreadsheets (
ods
)
As long as your files are simple enough, this library should just work. For anything else, please file an issue with a failing test or send a pull request!
Examples
Serde deserialization
It is as simple as:
use calamine::{open_workbook, Error, Xlsx, Reader, RangeDeserializerBuilder};
fn example() -> Result<(), Error> {
let path = format!("{}/tests/temperature.xlsx", env!("CARGO_MANIFEST_DIR"));
let mut workbook: Xlsx<_> = open_workbook(path)?;
let range = workbook.worksheet_range("Sheet1")?;
let mut iter = RangeDeserializerBuilder::new().from_range(&range)?;
if let Some(result) = iter.next() {
let (label, value): (String, f64) = result?;
assert_eq!(label, "celsius");
assert_eq!(value, 22.2222);
Ok(())
} else {
Err(From::from("expected at least one record but got none"))
}
}
Calamine provides helper functions to deal with invalid type values. For
instance, to deserialize a column which should contain floats but may also
contain invalid values (i.e. strings), you can use the
deserialize_as_f64_or_none
helper function with Serde's
deserialize_with
field attribute:
use calamine::{deserialize_as_f64_or_none, open_workbook, RangeDeserializerBuilder, Reader, Xlsx};
use serde::Deserialize;
#[derive(Deserialize)]
struct Record {
metric: String,
#[serde(deserialize_with = "deserialize_as_f64_or_none")]
value: Option<f64>,
}
fn main() -> Result<(), Box<dyn std::error::Error>> {
let path = format!("{}/tests/excel.xlsx", env!("CARGO_MANIFEST_DIR"));
let mut excel: Xlsx<_> = open_workbook(path)?;
let range = excel
.worksheet_range("Sheet1")
.map_err(|_| calamine::Error::Msg("Cannot find Sheet1"))?;
let iter_records =
RangeDeserializerBuilder::with_headers(&["metric", "value"]).from_range(&range)?;
for result in iter_records {
let record: Record = result?;
println!("metric={:?}, value={:?}", record.metric, record.value);
}
Ok(())
}
The
deserialize_as_f64_or_none
function discards all invalid values. If instead you would like to return them
as String
s, you can use the similar
deserialize_as_f64_or_string
function.
Reader: Simple
use calamine::{Reader, Xlsx, open_workbook};
let mut excel: Xlsx<_> = open_workbook("file.xlsx").unwrap();
if let Some(Ok(r)) = excel.worksheet_range("Sheet1") {
for row in r.rows() {
println!("row={:?}, row[0]={:?}", row, row[0]);
}
}
Reader: With header row
use calamine::{Reader, Xlsx, open_workbook};
let mut excel: Xlsx<_> = open_workbook("file.xlsx").unwrap();
let sheet1 = excel
.with_header_row(Some(3))
.worksheet_range("Sheet1")
.unwrap();
Note that xlsx
and xlsb
files support lazy loading, so specifying a
header row takes effect immediately when reading a sheet range.
In contrast, for xls
and ods
files, all sheets are loaded at once when
opening the workbook with default settings.
As a result, setting the header row only applies afterward and does not
provide any performance benefits.
Reader: More complex
Let's assume
- the file type (xls, xlsx ...) cannot be known at static time
- we need to get all data from the workbook
- we need to parse the vba
- we need to see the defined names
- and the formula!
use calamine::{Reader, open_workbook_auto, Xlsx, DataType};
// opens a new workbook
let path = ...; // we do not know the file type
let mut workbook = open_workbook_auto(path).expect("Cannot open file");
// Read whole worksheet data and provide some statistics
if let Some(Ok(range)) = workbook.worksheet_range("Sheet1") {
let total_cells = range.get_size().0 * range.get_size().1;
let non_empty_cells: usize = range.used_cells().count();
println!("Found {} cells in 'Sheet1', including {} non empty cells",
total_cells, non_empty_cells);
// alternatively, we can manually filter rows
assert_eq!(non_empty_cells, range.rows()
.flat_map(|r| r.iter().filter(|&c| c != &DataType::Empty)).count());
}
// Check if the workbook has a vba project
if let Some(Ok(mut vba)) = workbook.vba_project() {
let vba = vba.to_mut();
let module1 = vba.get_module("Module 1").unwrap();
println!("Module 1 code:");
println!("{}", module1);
for r in vba.get_references() {
if r.is_missing() {
println!("Reference {} is broken or not accessible", r.name);
}
}
}
// You can also get defined names definition (string representation only)
for name in workbook.defined_names() {
println!("name: {}, formula: {}", name.0, name.1);
}
// Now get all formula!
let sheets = workbook.sheet_names().to_owned();
for s in sheets {
println!("found {} formula in '{}'",
workbook
.worksheet_formula(&s)
.expect("sheet not found")
.expect("error while getting formula")
.rows().flat_map(|r| r.iter().filter(|f| !f.is_empty()))
.count(),
s);
}
Features
dates
: Add date related fn toDataType
.picture
: Extract picture data.
Others
Browse the examples directory.
Performance
As calamine
is readonly, the comparisons will only involve reading an excel xlsx
file and then iterating over the rows. Along with calamine
, three other libraries were chosen, from three different languages:
The benchmarks were done using this dataset, a 186MB
xlsx
file when the csv
is converted. The plotting data was gotten from the sysinfo
crate, at a sample interval of 200ms
. The program samples the reported values for the running process and records it.
The programs are all structured to follow the same constructs:
calamine
:
use calamine::{open_workbook, Reader, Xlsx};
fn main() {
// Open workbook
let mut excel: Xlsx<_> =
open_workbook("NYC_311_SR_2010-2020-sample-1M.xlsx").expect("failed to find file");
// Get worksheet
let sheet = excel
.worksheet_range("NYC_311_SR_2010-2020-sample-1M")
.unwrap()
.unwrap();
// iterate over rows
for _row in sheet.rows() {}
}
excelize
:
package main
import (
"fmt"
"github.com/xuri/excelize/v2"
)
func main() {
// Open workbook
file, err := excelize.OpenFile(`NYC_311_SR_2010-2020-sample-1M.xlsx`)
if err != nil {
fmt.Println(err)
return
}
defer func() {
// Close the spreadsheet.
if err := file.Close(); err != nil {
fmt.Println(err)
}
}()
// Select worksheet
rows, err := file.Rows("NYC_311_SR_2010-2020-sample-1M")
if err != nil {
fmt.Println(err)
return
}
// Iterate over rows
for rows.Next() {
}
}
ClosedXML
:
using ClosedXML.Excel;
internal class Program
{
private static void Main(string[] args)
{
// Open workbook
using var workbook = new XLWorkbook("NYC_311_SR_2010-2020-sample-1M.xlsx");
// Get Worksheet
// "NYC_311_SR_2010-2020-sample-1M"
var worksheet = workbook.Worksheet(1);
// Iterate over rows
foreach (var row in worksheet.Rows())
{
}
}
}
openpyxl
:
from openpyxl import load_workbook
# Open workbook
wb = load_workbook(
filename=r'NYC_311_SR_2010-2020-sample-1M.xlsx', read_only=True)
# Get worksheet
ws = wb['NYC_311_SR_2010-2020-sample-1M']
# Iterate over rows
for row in ws.rows:
_ = row
# Close the workbook after reading
wb.close()
Benchmarks
The benchmarking was done using hyperfine
with --warmup 3
on an AMD RYZEN 9 5900X @ 4.0GHz
running Windows 11
. Both calamine
and ClosedXML
were built in release mode.
0.22.1 calamine.exe
Time (mean ± σ): 25.278 s ± 0.424 s [User: 24.852 s, System: 0.470 s]
Range (min … max): 24.980 s … 26.369 s 10 runs
v2.8.0 excelize.exe
Time (mean ± σ): 44.254 s ± 0.574 s [User: 46.071 s, System: 7.754 s]
Range (min … max): 42.947 s … 44.911 s 10 runs
0.102.1 closedxml.exe
Time (mean ± σ): 178.343 s ± 3.673 s [User: 177.442 s, System: 2.612 s]
Range (min … max): 173.232 s … 185.086 s 10 runs
3.0.10 openpyxl.py
Time (mean ± σ): 238.554 s ± 1.062 s [User: 238.016 s, System: 0.661 s]
Range (min … max): 236.798 s … 240.167 s 10 runs
calamine
is 1.75x faster than excelize
, 7.05x faster than ClosedXML
, and 9.43x faster than openpyxl
.
The spreadsheet has a range of 1,000,001 rows and 41 columns, for a total of 41,000,041 cells in the range. Of those, 28,056,975 cells had values.
Going off of that number:
calamine
=> 1,122,279 cells per secondexcelize
=> 633,998 cells per secondClosedXML
=> 157,320 cells per secondopenpyxl
=> 117,612 cells per second
Plots
Disk Read
As stated, the filesize on disk is 186MB
:
calamine
=>186MB
ClosedXML
=>208MB
.openpyxl
=>192MB
.excelize
=>1.5GB
.
When asking one of the maintainers of excelize
, I got this response:
To avoid high memory usage for reading large files, this library allows user-specific UnzipXMLSizeLimit options when opening the workbook, to set the memory limit on the unzipping worksheet and shared string table in bytes, worksheet XML will be extracted to the system temporary directory when the file size is over this value, so you can see that data written in reading mode, and you can change the default for that to avoid this behavior.
- xuri
Disk Write
As seen in the previous section, excelize
is writting to disk to save memory. The others don't employ that kind of mechanism.
Memory
[!NOTE]
ClosedXML
was reporting a constant2.5TB
of virtual memory usage, so it was excluded from the chart.
The stepping and falling for calamine
is from the grows of Vec
s and the freeing of memory right after, with the memory usage dropping down again. The sudden jump at the end is when the sheet is being read into memory. The others, being garbage collected, have a more linear climb all the way through.
CPU
Very noisy chart, but excelize
's spikes must be from the GC?
Unsupported
Many (most) part of the specifications are not implemented, the focus has been put on reading cell values and vba code.
The main unsupported items are:
- no support for writing excel files, this is a read-only library
- no support for reading extra contents, such as formatting, excel parameter, encrypted components etc ...
- no support for reading VB for opendocuments
Credits
Thanks to xlsx-js developers! This library is by far the simplest open source implementation I could find and helps making sense out of official documentation.
Thanks also to all the contributors!
License
MIT
Dependencies
~8MB
~201K SLoC