Supported Formats

Extra notes about formats

tsv

• loader-specific options
  • delimiter (default: '\t') field delimiter to use for tsv/usv filetype
  • row_delimiter (default: '\n') row delimiter to use for tsv/usv filetype
  • tsv_safe_newline (default: '\u001e') replacement for newline character when saving to tsv
  • tsv_safe_tab (default: '\u001f') replacement for tab character when saving to tsv

csv

• loader-specific options
  • csv_dialect (default: 'excel') dialect passed to csv.reader
    • Accepted dialects are excel-tab, unix, and excel.
  • csv_delimiter (default: ',') delimiter passed to csv.reader
  • csv_quotechar (default: '"') quotechar passed to csv.reader
  • csv_skipinitialspace (default: True) skipinitialspace passed to csv.reader
  • csv_escapechar (default: None) escapechar passed to csv.reader
  • csv_lineterminator (default: '\r\n') lineterminator passed to csv.writer
  • safety_first (default: False) sanitize input/output to handle edge cases, with a performance cost

fixed

• loader-specific options
  • fixed_rows (default: 1000) number of rows to detect fixed width columns from
  • fixed_maxcols (default: 0) max number of fixed-width columns to create (0 is no max)

json

• loader-specific options
  • json_indent (default: None) indent to use when saving json
  • json_sort_keys (default: False) sort object keys when saving to json
  • default_colname (default: '') column name to use for non-dict rows
• Cells containing lists (e.g. [3]) or dicts (e.g. {3}) can be expanded into new columns with ( and unexpanded with ).
• All expanded subcolumns must be closed (with )) to retain the same structure.

xml

• v show only columns in current row attributes
• za add column for xml attributes

pcap

• loader-specific options
  • pcap_internet (default: 'n') (y/s/n) if save_dot includes all internet hosts separately (y), combined (s), or does not include the internet (n)

postgres

• loader-specific options
  • postgres_schema (default: 'public') the desired schema for the Postgres database
• vd postgres://username:password@hostname:port/database opens a connection to the given postgres database.

using VisiData as a pager within psql

In psql:

\pset format csv
\pset pager always
\setenv PSQL_PAGER 'vd -f csv'
\pset pager_min_lines

sqlite

• supports saving for CREATE/INSERT (not wholesale updates)
• z Ctrl+S to commit any add-row/edit-cell/delete-row

mysql

• loader-specific requirements
  • working mysql / mariadb installation or at least the libmysqlclient-dev package (ubuntu; name might be different on other platforms)
  • mysqlclient python module in path or virtual environment (pip install mysqlclient)
• vd mysql://username:password@hostname:port/database opens a connection to the given mysql / mariadb database.

html

• loader-specific options
  • html_title (default: '<h2>{sheet.name}</h2>') table header when saving to html
• load all <table>s in a web page as VisiData sheets.

shp

• Can be edited in raw data form. Images can be plotted with . (dot).
• .shp files can be saved as geoJSON.

mbtiles

• Can be edited in raw data form. Images can be plotted with . (dot).

png

• Can be edited in raw data form. Images can be plotted with . (dot).

ttf

• Can be edited in raw data form. Images can be plotted with . (dot).

pandas

VisiData has an adapter for pandas. To load a file format which is supported by pandas, pass -f pandas data.foo. This will call pandas.read_foo().

For example:

vd -f pandas data.parquet

loads a parquet file. When using the pandas loader, the .fileformat file extension is mandatory.

To load a hierarchy of parquet files located in folder data/, run

vd -f parquet data/

or rename the directory to data.parquet and run

vd data.parquet -f pandas

This should similarly work for any format that has a pandas.read_format() function.

vd

• Command history log format for a VisiData session.
• Ctrl+D to save the current session's CommandLog.