csv2html

Crates.iocsv2html
lib.rscsv2html
version3.1.1
sourcesrc
created_at2023-07-12 10:35:34.635326
updated_at2024-05-24 12:51:19.036427
descriptionConvert CSV files to HTML tables
homepagehttps://github.com/dbohdan/csv2html
repositoryhttps://github.com/dbohdan/csv2html
max_upload_size
id914423
size50,460
D. Bohdan (dbohdan)

documentation

README

csv2html

This command-line utility converts CSV files to HTML tables and complete HTML documents. It requires Rust 1.74 or later to build. By default, it uses the first row of the CSV file as the header of the HTML table.

The original Python version of csv2html is preserved in the branch python.

Installation

Prebuilt Linux and Windows binaries are available. They are attached to releases on the "Releases" page.

Installing with Cargo

cargo install csv2html

Building on Debian and Ubuntu

Follow the instructions to build a static Linux binary of csv2html from the source code on recent Debian and Ubuntu.

1. Install Rustup. Through Rustup add the stable musl libc target for your CPU.

rustup target add x86_64-unknown-linux-musl

2. Install the build and test dependencies.

sudo apt install build-essential musl-tools
cargo install just

3. Clone this repository. Build the binary.

git clone https://github.com/dbohdan/csv2html
cd csv2html
just test
just release-linux

Cross-compiling for Windows

Follow the instructions to build a 32-bit Windows binary of csv2html on recent Debian and Ubuntu.

1. Install Rustup. Through Rustup add the i686 GNU ABI Windows target.

rustup target add i686-pc-windows-gnu

2. Install the build dependencies.

sudo apt install build-essential mingw-w64
cargo install just

3. Configure Cargo for cross-compilation. Add the following in ~/.cargo/config.

[target.i686-pc-windows-gnu]
linker = "/usr/bin/i686-w64-mingw32-gcc"

4. Clone this repository. Build the binary.

git clone https://github.com/dbohdan/csv2html
cd csv2html
just release-windows

Command-line arguments

Convert CSV files to HTML tables

Usage: csv2html [OPTIONS] [input]

Arguments:
  [input]  Input file

Options:
  -o, --output <OUTPUT>    Output file
  -t, --title <TITLE>      HTML document title
  -d, --delimiter <DELIM>  Field delimiter character for CSV (',' by default)
  -s, --start <N>          Skip the first N-1 rows; start at row N
  -r, --renumber           Replace the first column with row numbers
  -n, --no-header          Do not use the first row of the input as the header
  -c, --complete-document  Output a complete HTML document instead of only a
table
      --table <ATTRS>      HTML attributes for the tag <table> (e.g., --table
'foo="bar" baz' results in the output <table foo="bar" baz>...</table>); it is
up to the user to ensure the result is valid HTML
      --tr <ATTRS>         Attributes for <tr>
      --th <ATTRS>         Attributes for <th>
      --td <ATTRS>         Attributes for <td>
  -h, --help               Print help
  -V, --version            Print version

Use examples

The following command reads the data in test/test.csv and writes an HTML table that corresponds to the data to test.html:

csv2html -o test.html tests/test.csv

The following command takes semicolon-delimited data from pub.csv, starting with row 267. The first column of the table is replaced with the row number starting at 1 (except in the header row, which is untouched). The output is redirected to the file pub.html.

csv2html pub.csv -d \; -r -s 267 > pub.html

The same as above, but the output is a full HTML document instead of just the markup for the table:

csv2html pub.csv -d \; -r -s 267 -c > pub.html

If the input file is tab-delimited, use \t as the deliminter argument.

# POSIX.
csv2html --delimiter '\t' tests/test.tsv
rem Windows.
csv2html-win32.exe --delimiter \t tests/test.tsv

\t is the only backslash escape sequence that is implemented.

License

Three-clause ("new" or "revised") BSD. See the file LICENSE.

Commit count: 114

cargo fmt