Performance degradation in sa.annotated.pandas

[kbrodt]

Hello,

I launched a simple comparison of apply between pandas and sa.annotated.pandas and got performance 0.018s in the first case and 12.47s in the second one. Also I cannot print annotated dataframe. What I do uncorrect here?

To reproduce:

import time


def test_new_apply(is_sa=False):
 if is_sa:
 import sa.annotated.pandas as pd
 else:
 import pandas as pd

 a = pd.DataFrame(list(range(10 ** 5)), columns=['x'])
 if is_sa:
 # print(a.head()) # <- raises AttributeError: 'Operation' object has no attribute 'max'
 pass
 else:
 print(a.head())
 start = time.time()
 a['xp1'] = a['x'].apply(lambda x: x + 1)
 print(time.time() - start)


test_new_apply(is_sa=False)
test_new_apply(is_sa=True)

Environment:

Python 3.6.8
[GCC 8.3.0] on linux

pandas==0.25.2
sa==0.0.4

[kbrodt]

Probably It is due to non annotated lambda.