CleverCSV provides a drop-in replacement for the Python csv package with improved dialect detection for messy CSV files. It also provides a handy command line tool that can standardize a messy file or generate Python code to import it.

Useful links:

• CleverCSV on Github
• CleverCSV on PyPI
• Demo of CleverCSV on Binder (interactive!)
• Research Paper on CSV dialect detection (PDF)
• Reproducible Research Repo
• Blog post on messy CSV files^NEW!

Introduction

• CSV files are awesome! They are lightweight, easy to share, human-readable, version-controllable, and supported by many systems and tools!
• CSV files are terrible! They can have many different formats, multiple tables, headers or no headers, escape characters, and there's no support for recording metadata!

CleverCSV is a Python package that aims to solve some of the pain points of CSV files, while maintaining many of the good things. The package automatically detects (with high accuracy) the format (dialect) of CSV files, thus making it easier to simply point to a CSV file and load it, without the need for human inspection. In the future, we hope to solve some of the other issues of CSV files too.

CleverCSV is based on science. We investigated thousands of real-world CSV files to find a robust way to automatically detect the dialect of a file. This may seem like an easy problem, but to a computer a CSV file is simply a long string, and every dialect will give you some table. In CleverCSV we use a technique based on the patterns of row lengths of the parsed file and the data type of the resulting cells. With our method we achieve a 97% accuracy for dialect detection, with a 21% improvement on non-standard (messy) CSV files compared to the Python standard library.

We think this kind of work can be very valuable for working data scientists and programmers and we hope that you find CleverCSV useful (if there's a problem, please open an issue!) Since the academic world counts citations, please cite CleverCSV if you use the package. Here's a BibTeX entry you can use:

@article{van2019wrangling,
        title = {Wrangling Messy {CSV} Files by Detecting Row and Type Patterns},
        author = {{van den Burg}, G. J. J. and Naz{\'a}bal, A. and Sutton, C.},
        journal = {Data Mining and Knowledge Discovery},
        year = {2019},
        volume = {33},
        number = {6},
        pages = {1799--1820},
        issn = {1573-756X},
        doi = {10.1007/s10618-019-00646-y},
}

And of course, if you like the package please spread the word! You can do this by Tweeting about it (#CleverCSV) or clicking the ⭐️ on GitHub!

Installation

The package is available on PyPI:

$ pip install clevercsv

Usage

CleverCSV consists of a Python library and a command line tool called clevercsv.

Library

We designed CleverCSV to provide a drop-in replacement for the built-in CSV module, with some useful functionality added to it. Therefore, if you simply want to replace the builtin CSV module with CleverCSV, you can import CleverCSV as follows, and use it as you would use the builtin csv module.

import clevercsv

CleverCSV provides an improved version of the dialect sniffer in the CSV module, but it also adds some useful wrapper functions. These functions automatically detect the dialect and aim to make working with CSV files easier. We currently have the following helper functions:

• detect_dialect: takes a path to a CSV file and returns the detected dialect
• read_csv: automatically detects the dialect and encoding of the file, and returns the data as a list of rows. A version that returns a generator is also available: stream_csv
• csv2df: detects the dialect and encoding of the file and then uses Pandas to read the CSV into a DataFrame.
• write_table: write a table (a list of lists) to a file using the RFC-4180 dialect.

Of course, you can also use the traditional way of loading a CSV file, as in the Python CSV module:

# importing this way makes it easy to port existing code to CleverCSV!
import clevercsv as csv

with open("data.csv", "r", newline="") as fp:
  # you can use verbose=True to see what CleverCSV does:
  dialect = csv.Sniffer().sniff(fid.read(), verbose=False)
  fp.seek(0)
  reader = csv.reader(fp, dialect)
  rows = list(reader)

That's the basics! If you want more details, you can look at the code of the package, the test suite, or the API documentation.

Command-Line Tool

The clevercsv command line application has a number of handy features to make working with CSV files easier. For instance, it can be used to view a CSV file on the command line while automatically detecting the dialect. It can also generate Python code for importing data from a file with the correct dialect. The full help text is as follows:

USAGE
  clevercsv [-h] [-v] [-V] <command> [<arg1>] ... [<argN>]

ARGUMENTS
  <command>       The command to execute
  <arg>           The arguments of the command

GLOBAL OPTIONS
  -h (--help)     Display this help message.
  -v (--verbose)  Enable verbose mode.
  -V (--version)  Display the application version.

AVAILABLE COMMANDS
  code            Generate Python code for importing the CSV file.
  detect          Detect the dialect of a CSV file
  help            Display the manual of a command
  standardize     Convert a CSV file to one that conforms to RFC-4180.
  view            View the CSV file on the command line using TabView

Each of the commands has further options (for instance, the code command can generate code for importing a Pandas DataFrame). Use clevercsv help <command> for more information. Below are some examples for each command:

Code

Code generation is useful when you don't want to detect the dialect of the same file over and over again. You simply run the following command and copy the generated code to a Python script!

$ clevercsv code imdb.csv

# Code generated with CleverCSV

import clevercsv

with open("imdb.csv", "r", newline="", encoding="utf-8") as fp:
    reader = clevercsv.reader(fp, delimiter=",", quotechar="", escapechar="\\")
    rows = list(reader)

We also have a version that reads a Pandas dataframe:

$ clevercsv code --pandas imdb.csv

# Code generated with CleverCSV

import clevercsv

df = clevercsv.csv2df("imdb.csv", delimiter=",", quotechar="", escapechar="\\")

Detect

Detection is useful when you only want to know the dialect.

$ clevercsv detect imdb.csv
Detected: SimpleDialect(',', '', '\\')

The --plain flag gives the components of the dialect on separate lines, which makes combining it with grep easier.

$ clevercsv detect --plain imdb.csv
delimiter = ,
quotechar =
escapechar = \

Standardize

Use the standardize command when you want to rewrite a file using the RFC-4180 standard:

$ clevercsv standardize --output imdb_standard.csv imdb.csv

In this particular example the use of the escape character is replaced by using quotes.

View

This command allows you to view the file in the terminal. The dialect is of course detected using CleverCSV! Both this command and the standardize command support the --transpose flag, if you want to transpose the file before viewing or saving:

$ clevercsv view --transpose imdb.csv

Contributing

If you want to encourage development of CleverCSV, the best thing to do now is to spread the word!

If you encounter an issue in CleverCSV, please open an issue or submit a pull request. Don't hesitate, you're helping to make this project better! If GitHub's not your thing but you still want to contact us, you can send an email to gertjanvandenburg at gmail dot com instead.

Note that all contributions to the project must adhere to the Code of Conduct.

The CleverCSV package was originally written by Gertjan van den Burg and came out of scientific research on wrangling messy CSV files by Gertjan van den Burg, Alfredo Nazabal, and Charles Sutton.

Notes

License: MIT (see LICENSE file).

Copyright (c) 2019-2020 The Alan Turing Institute.