Closed hagenw closed 2 months ago
If we request first the existing schemes, it works as expected:
>>> db.get("emotion", additional_schemes=["speaker", "sex", "bla"], tables="emotion.categories.test.gold_standard")
emotion speaker sex bla
file
1048/1048_IEO_NEU_XX.wav neutral 1048 male NaN
1048/1048_IEO_HAP_LO.wav neutral 1048 male NaN
1048/1048_IEO_HAP_MD.wav neutral 1048 male NaN
1048/1048_IEO_HAP_HI.wav neutral 1048 male NaN
1048/1048_IEO_SAD_LO.wav neutral 1048 male NaN
... ... ... ... ...
1091/1091_WSI_HAP_XX.wav neutral 1091 female NaN
1091/1091_WSI_SAD_XX.wav neutral 1091 female NaN
1091/1091_WSI_ANG_XX.wav anger 1091 female NaN
1091/1091_WSI_FEA_XX.wav sadness 1091 female NaN
1091/1091_WSI_DIS_XX.wav neutral 1091 female NaN
[1392 rows x 4 columns]
But from the point where we request a non-existing scheme, everything else is also set to NaN
:
>>> db.get("emotion", additional_schemes=["bla", "speaker", "sex"], tables="emotion.categories.test.gold_standard")
emotion bla speaker sex
file
1048/1048_IEO_NEU_XX.wav neutral NaN NaN NaN
1048/1048_IEO_HAP_LO.wav neutral NaN NaN NaN
1048/1048_IEO_HAP_MD.wav neutral NaN NaN NaN
1048/1048_IEO_HAP_HI.wav neutral NaN NaN NaN
1048/1048_IEO_SAD_LO.wav neutral NaN NaN NaN
... ... ... ... ...
1091/1091_WSI_HAP_XX.wav neutral NaN NaN NaN
1091/1091_WSI_SAD_XX.wav neutral NaN NaN NaN
1091/1091_WSI_ANG_XX.wav anger NaN NaN NaN
1091/1091_WSI_FEA_XX.wav sadness NaN NaN NaN
1091/1091_WSI_DIS_XX.wav neutral NaN NaN NaN
[1392 rows x 4 columns]
For the following example, we need the
crema-d
dataset:The dataset has a
sex
scheme, that we can request as additional scheme:But when requesting the same scheme, but another in addition, we get nothing back: