Last updated 12 October 2026
Data integrity
What is preserved
For delimited text and JSON, Polypress preserves the column names and their order, the row order and the text of every cell. Formatting that does not change the content, such as quotation marks around fields and line endings, may differ from the original file. Stata, SPSS and SAS files are restored byte for byte.
Verification
By default, each archive is decompressed in memory and compared with the input, cell by cell,
before it is written. The --no-verify option disables this check.
Rows of unequal length
- A row with fewer cells than the header is padded with empty cells.
- A row with more cells than the header is accepted only if the additional cells are empty. Otherwise the file is rejected and the row is reported.
Text encoding
Input files are read as UTF-8, or according to their byte order mark if one is present. Polypress rejects a file that is not valid UTF-8 and reports the position of the first invalid byte. To read such a file, specify its encoding:
polypress compress survey.csv --encoding latin-1
Determinism and versions
With the same version of Polypress, the same input file produces an identical archive on any
machine and with any number of threads. Each archive records the version of Polypress that
created it, which polypress info reports. The program reports its version as 1.00;
the corresponding Python package is version 1.0.0.