Keyword export ordering

Feature Requests
, ,
1

Dear Jason,
Sorry, I am getting back into this only now.

  • Regarding keyword export:
    The Encounter.mediaAssetX.keywordX are exported, but they are not displayed in the same column in the Excel spreadsheet.
    For ex, if you look at the export file I sent (Export Project data
    /3?u=violaine.dulau ), the column BS include the keywords “feature” and “distinctivness”. This is because sometime the keyword “Profile photo” is used and sometime not, which lead to a shift in the colmuns used to exprot the data. We would like to have each labelled keywords (feature, distinctiivness, quality, fluke type) exported in a single unique column (for ex all features in column BS", all distinctiveness keywords in column BT, ect…). Otherwise it is very difficult to filter the data, and use the file for further analysis.
    Ideally, we would like to have the name of the labelled keyword as the name of the column, and the values provided in that column (fo ex, have one exported column named “Feature” and the values (fluke, dorsal fin) in the column.
    If it is not possible, one solution would be to export an NA or NULL value when there is no value for the labelled keywords (including profile photo) so that the keywords values are all exported in the same column.
    Thanks,
    Violaine
2

Dear Jason,
Sorry, I am getting back into this only now.

  • Regarding keyword export:
    The Encounter.mediaAssetX.keywordX are exported, but they are not displayed in the same column in the Excel spreadsheet.
    if you look at the export file
3

Thanks @violaine.dulau

We will look at this as a feature request for a change.

4

Tracking this as ticket WB-1996 and exploring a fix.

5

Hi @violaine.dulau

I may have a compromise we can work with.

I have modified the export to list each labeled keyword by its name, which matches the bulk import format now too. So distinctiveness is now listed in column Encounter.mediaAsset0.distinctiveness with the value 0 listed. I cannot guarantee column number/letters and order because each image may have multiple keywords and therefore columns Encounter.mediaAsset0.keyword0, Encounter.mediaAsset0.keyword1, etc. can push out Encounter.mediaAsset0.distinctiveness since labeled keywords get added after the normal keywords. However, almost every analytics package can pivot on column name, which was not possible when labeled keywords were part of keywords previously and did not have distinct column headers.

If it is not possible, one solution would be to export an NA or NULL value when there is no value for the labelled keywords (including profile photo) so that the keywords values are all exported in the same column.

This would unfortunately bloat the spreadsheet as every possible keyword would need to be listed for every possible image. We avoid “NA” and “NULL” because they can be interpreted as strings in analytics packages, whereas the empty column is unambiguslously null or empty.

Can you please try a few exports and see if you can work with the current format?

Thank you,
Jason