Welcome to Gtable’s documentation!¶
Gtable (source) is a container for
tabular or tabular-like data designed with speed in
mind. It is very similar to pandas, and it relies
on many of its capabilities. It tries to improve one some aspects of
using pandas data types pandas.Series
and
pandas.DataFrame
as containers for simple computations, but it is
not a replacement for them.
- It tries to reduce the overhead for column access, creation, and concatenation.
- It supports sparse data with bitmap indexing.
- It truly handles NaNs, making a difference between a NaN and a NA in its internal representation.
- It provides fast transformations (filling NA values, filtering, joining…)
It also relies heavily on numpy. You can consider gtable as a thin layer over numpy arrays.
You can install gtable with a simple pip install gtable
.
Gtable is an open-source project released under a BSD 3-Clause license. You can find a copy of the license in this document