- Remove the low_memory=True option. This change should improve the performance and identify the column types based on shmessy sample instead of pandas dtypes-guess mechanism
1.2.6
- New datetime format: `%d/%m/%Y %H:%M` - Column number limitation (500 by default)
1.2.5
- Support "0"/"1" (As string) identification for boolean type - Improved PB tests - Fix readme pypi version badge
1.2.4
- Fallback to null option - Convert the problematic data to null empty value instead of raising an exception. - Fix boolean type - In some cases not negative data has recognized as FALSE value.
1.2.3
- Fix: Remove space from the delimiters options, Might cause some issues while reading CSV files with spaces in the data. - Add black badge to the readme file
1.2.2
- Add option to ignore virtual types (In order to improve the execution time) - Add casting by entire column to improve DF fix duration - Disable dynamic type discovery to improve duration