Skip to content

Latest commit

 

History

History
150 lines (116 loc) · 5.64 KB

README.md

File metadata and controls

150 lines (116 loc) · 5.64 KB
Raw

pd: Command-line Interface to Pandas for Dealing with Tabular Data

This is a little python tool to quickly handle tabular data in the command line, like this:

cat example.csv | pd -q "x > 30" --sort z > output.csv

It supports pandas queries, some basic column manipulation (renaming/reordering/...), different input and output formats (csv, json, latex, html) and sorting.

This tool should not be considered stable or complete in any kind yet and will likely be inefficient when dealing with very large csv files. It is rather meant for dealing with small and medium-sized tabular data on the command line or in scripts (e.g. quickly converting a csv to a latex table or sorting and filtering values).

Installation

Manual installation

Requires python >= 3.6 and pip. Clone the git repo and install the requirements with pip install --user -r requirements.txt.

Run pd with ./pd.py. Below it is assumed that pd is an executable which links to the script.

TODO: proper installation setup with pip

Docker

There is a docker image available if you have docker installed:

docker run dhelmr/pd --help

Here is an example for reading in csv from stdin and converting it to json:

cat examples/example.csv | docker run -i dhelmr/pd --to json

Note that the -i and -o options won't work out-of-the-box when using the docker image. You'll need to mount the input and output files to the docker container:

docker run -v $(pwd)/examples/example.csv:/input dhelmr/pd -i /input --to json

Examples

Read a table from json, sort records by a column and store as json

pd --from json -i example.json --sort z --to json > sorted_by_z.json

convert a csv to a latex table and highlight the maximum number of some columns ('x' and 'z' here):

pd --from csv -i example.csv --to latex --latex-max-bold x z -o examples.tex

list columns of a csv

pd -i example.csv --columns

rename column:

pd -i example.csv --rename y::new_column_name > output.csv

Filter columns and print as a nice table

pd -i example.csv --only x y --pretty

Help

All CLI arguments:

usage: pd [-h] [--query QUERY] [--input INPUT] [--output OUTPUT] [--pretty]
          [--pretty-short] [--transpose] [--sort SORT [SORT ...]]
          [--descending] [--drop DROP [DROP ...]] [--with-index] [--to TO]
          [--from FROM_FORMAT] [--float-format FLOAT_FORMAT]
          [--move MOVE [MOVE ...]] [--rename RENAME [RENAME ...]] [--sep SEP]
          [--latex-max-bold LATEX_MAX_BOLD [LATEX_MAX_BOLD ...]]
          [--replace-nan REPLACE_NAN]
          [--json-out-orient {split,records,index,columns,values,table}]
          [--json-in-orient {split,records,index,columns,values,table}]
          [--only ONLY [ONLY ...]] [--columns] [-n] [--group-by GROUP_BY]
          [--group-by-max GROUP_BY_MAX] [--unique UNIQUE]

options:
  -h, --help            show this help message and exit
  --query QUERY, -q QUERY
                        Pandas query
  --input INPUT, -i INPUT
                        Input file
  --output OUTPUT, -o OUTPUT
                        Output file, if not set the output is printed to
                        stdout
  --pretty, -p
  --pretty-short, -pp
  --transpose, -t
  --sort SORT [SORT ...], -s SORT [SORT ...]
                        Sort by column(s)
  --descending          Sort descending
  --drop DROP [DROP ...]
                        Drop column(s)
  --with-index          Specify for outputing the index column
  --to TO               Specify the output format. Choose from: ['csv',
                        'latex', 'json', 'html']
  --from FROM_FORMAT    Specify the input format. Choose from: ['csv', 'json']
  --float-format FLOAT_FORMAT
                        Specify the float format in the output, e.g. '%.4f'
  --move MOVE [MOVE ...]
                        Moves column X right before after column Y; format:
                        'X::Y'
  --rename RENAME [RENAME ...]
                        Renames columns, format: 'old name::new name'
  --sep SEP             Column seperator
  --latex-max-bold LATEX_MAX_BOLD [LATEX_MAX_BOLD ...]
                        Marks the maximum value of each column in the latex
                        output bold.
  --replace-nan REPLACE_NAN
                        Replace all NaN values with another value
  --json-out-orient {split,records,index,columns,values,table}
                        Specifies the json input orient. Choose one of
                        ['split', 'records', 'index', 'columns', 'values',
                        'table']. See https://pandas.pydata.org/pandas-docs/st
                        able/reference/api/pandas.Series.to_json.html?highligh
                        t=to_json#pandas.Series.to_json
  --json-in-orient {split,records,index,columns,values,table}
                        Specifies the json output orient. Choose one of
                        ['split', 'records', 'index', 'columns', 'values',
                        'table']. See https://pandas.pydata.org/pandas-docs/st
                        able/reference/api/pandas.Series.to_json.html?highligh
                        t=to_json#pandas.Series.to_json
  --only ONLY [ONLY ...]
  --columns             Prints the columns
  -n, --nrows           Prints the number of rows
  --group-by GROUP_BY   Groups by the specified column.
  --group-by-max GROUP_BY_MAX
                        Determines the maximum column value of a group by.
  --unique UNIQUE       Prints unique values of the specified column.

License

Licensed under GPLv3.