.query on large DataFrame yields "None of [Int64Index([...] dtype='int64')] are in the [columns]"

[modal-python]

.query on large DataFrame yields "None of [Int64Index([...] dtype='int64')] are in the [columns]"

[cosmic-cortex]

Unfortunately, modAL currently does not support pandas support for technical reasons. I have detailed the problem in #20, but the gist is, numpy arrays use row first indexing, while pandas DataFrames use column first by default. This led to tehcnical difficulties which I was unable to implement a proper solution. Of course there would be workarounds, but I did not come up with a robust solution.

[srowen]

Oh I see, didn't realize as it happened to mostly work fine with DataFrames.
Isn't .iloc how you'd do row indexing in pandas? You could selectively calls this if it's a DataFrame.
For my use case numpy arrays would work fine too though.

[cosmic-cortex]

Yes, .iloc would work, but I didn't want to selectively call X[idx] or X.iloc[idx] based on the datatype. That solution would be very hard to extend for a third datatype. A possible solution would be to introduce a wrapper class modALinput which would handle all the supported datatypes.

[vidushi-chouksey]

If I use .loc, then facing the error as stated below:
KeyError: "None of [Int64Index([ 160, 161, 162, 163, 164, 165, 166, 167, 168, 169,\n ...\n 1590, 1591, 1592, 1593, 1594, 1595, 1596, 1597, 1598, 1599],\n dtype='int64', name='', length=1440)] are in the [index]"

whereas; if I use .iloc, then facing the following error:
IndexError: positional indexers are out-of-bounds

Could you please let me know how to approach this.