pdlog 0.1.0.post0

Logging for pandas dataframes

pdlog

pdlog provides logging for pandas dataframes, to better enable you to monitor and debug your data pipelines.

For example:

>>> import pdlog
>>> df = df.log.dropna()
2020-05-26 20:55:30,049 INFO <pdlog> dropna: dropped 1 row (17%), 5 rows remaining

Example

The above assumes that the logging module has been configured and that data has been loaded into a pandas DataFrame. Let's walk through those steps with a simple example.

1. Configure logging:

   >>> import logging
   >>> fmt = "{asctime} {levelname} <{name}> {message}"
   >>> logging.basicConfig(format=fmt, style="{", level=logging.INFO)
   

2. Load data into a pandas.DataFrame:

   >>> import pandas as pd
   >>> df = pd.DataFrame([0, 1, 2, None, 4])
   >>> df.head()
        0
   0  0.0
   1  1.0
   2  2.0
   3  NaN
   4  4.0
   

3. Importing pdlog and call a method under the log accessor:

   >>> import pdlog
   >>> df = df.log.dropna()
   2020-05-26 20:55:30,049 INFO <pdlog> dropna: dropped 1 row (17%), 5 rows remaining
   

Supported methods

pdlog currently supports the following pandas.DataFrame methods:

• Filter rows and select columns:
  • drop_duplicates
  • drop
  • dropna
  • head
  • query
  • sample
  • tail
• (Re-)set indexes:
  • reset_index
  • set_index
• Rename indexes:
  • rename
• Reshape:
  • melt
  • pivot
• Impute:
  • bfill
  • ffill
  • fillna

Related Work

pandas-log

pandas-log is aimed at interactive usage. Its messages are friendlier and more verbose than pdlog aims to be. Ideally, each pdlog message should be a single line of dense information to help you understand whether your production code is doing what you think it is, while not overflowing your logs. These don't tend to make particularly friendly messages.

tidylog

pdlog can be considered a port of tidylog (R package) to pandas. Their goals align with ours, and we think they've done a great job at reaching those goals.