I read your post "Efficient Tabular Storage" and decided to to give Castra a whirl. As my luck would have it, I got an error on the very first data set I tried.
The problem is that the dataframe has a column that is labeled "." (period, without quotes) and, of course, on *nix this is an invalid name for regular files. I don't know if/how you would want to fix this. I've written a test case nontheless.
def test_column_with_period():
df = pd.DataFrame({'x': [10, 20],
'.': [10., 20.]},
columns=['x', '.'],
index=[10, 20])
with Castra(template=df) as c:
c.extend(df)
Which yields:
> with open(filename, 'wb') as fp:
E IOError: [Errno 21] Is a directory: '/tmp/castra-uUz3Gr/10--20/.'
Hi @enku thanks for trying it out. I hope that it can (eventually) be useful for you. There is a fix for this immediate issue in #42. It could be improved though if you have suggestions. (thanks also, for the test.)
I read your post "Efficient Tabular Storage" and decided to to give Castra a whirl. As my luck would have it, I got an error on the very first data set I tried.
The problem is that the dataframe has a column that is labeled "." (period, without quotes) and, of course, on *nix this is an invalid name for regular files. I don't know if/how you would want to fix this. I've written a test case nontheless.
Which yields: