pandas 1.0 was released 2 years ago, and we still have multiple checks for pandas versions for even 0.2
For context:
v0.20.3 (July 7, 2017)
v0.23.0 (May 15, 2018)
1.0.0 (January 29, 2020)
Latest: 1.4.0 (Jan 22, 2022)
I think requiring pandas >=1.0 is reasonable because 1.0 dropped many of announced deprecations, and was released long enough ago. This would simplify code by allowing to drop checking of multiple pandas versions. Besides, we do not really test pandapower with older pandas versions, which is also a potential risk.
jkisse
commented
2 months ago
pandas 1.0 was released 2 years ago, and we still have multiple checks for pandas versions for even 0.2
For context:
v0.20.3 (July 7, 2017)
v0.23.0 (May 15, 2018)
1.0.0 (January 29, 2020)
Latest: 1.4.0 (Jan 22, 2022)
I think requiring pandas >=1.0 is reasonable because 1.0 dropped many of announced deprecations, and was released long enough ago. This would simplify code by allowing to drop checking of multiple pandas versions. Besides, we do not really test pandapower with older pandas versions, which is also a potential risk.