Build fail: KeyError 'MIM Number' #106
Conversation
joeflack4
force-pushed
the
bugfix-keyerr-mimnum
branch
from
834612c
to
ffe6b70
Compare
| try: | ||
| df = pd.read_csv(input_path, delimiter='\t', comment='#').fillna('') | ||
| df[mim_col] = df[mim_col].astype(int) # these were being read as floats | ||
| # TODO: Need a better solution than this. Which should be: When these files are downloaded, should uncomment header |
There was a problem hiding this comment.
The issue
I'm not sure why this suddenly started happening, where the header is missing from the files. This is what is causing the KeyError causing the action to fail.
The solution
The best way to fix this would be to do things at download time for all the files. But in the interest of time, I fixed it at the point of failure.
Need a better solution than this. Which should be: When these files are downloaded, should uncomment header
There was a problem hiding this comment.
What issue is this solving? Sorry I was in a big hole, but getting back to life now!
There was a problem hiding this comment.
That's all right. Thanks for checking in on this. I wanted more time to look more deeply into the root cause and fix the issue upstream in a better way, but basically I did a quick fix.
Basically what's happening is that when the TSVs are downloaded, there are a few comment lines, including the header is commented out. So we do a special operation to uncomment that header . For some reason it suddenly it wasn't getting uncommented
Updates