Metadata index can't handle value 'NA'

[thermokarst]

The pandas parsing interprets 'NA' as np.nan, causing an index validation error. I suspect this also happens if a column name is 'NA', too.

This recently came up on the forum.

[ElDeveloper]

I've faced this issue, and solved it by setting the na_values=[], keep_default_na=False arguments.

[antgonza]

Suggest taking a look at the Qiita load_template_to_dataframe function. It literally took us years to get it right (with multiple users updating templates) and we haven't had an issue in a while.

[thermokarst]

This came up on the forum.

[jairideout]

forum x-ref

[jairideout]

Fixed in #359; confirmed that IDs and columns can be named NA, which is no longer interpreted as a missing value.