Wrong data_id when importing dataset

[ClementLeBihan]

Hi, When importing big dataset with annotations, it happen that some frame lose its annotation.

When opening the data_annotation_object, I can see the data_id field that is weird for the corresponding frame that hasn't correctly loaded the results.

Do you know this issue ?

[ClementLeBihan]

I investigated more : 1 - The problem occurs randomly, 2 - When it happen for a frame, in saveBatchDataResult the dataIdMap missmatch the TempDataId of dataAnnotationObjectBOList : For exemple the dataIdMap will be {-696053440268826030=31, -1259394534141999808=34, 3904700820343092532=30, -7117869251104120990=32, 1357986784990387562=33} while TempDataId of dataAnnotationObjectBOList are [2924171183170165039, -696053440268826030, -7117869251104120990, 1357986784990387562 , -1259394534141999808] We can see that the first one doesn't have key in dataIdMap 3 - When I set parseExecutorService corePoolSize to 1, I don't have the bug.