Feature Type

  • [x] Adding new functionality to pandas

  • [ ] Changing existing functionality in pandas

  • [ ] Removing existing functionality in pandas

Problem Description

When using pd.HDFStore methods directly the store often stays open. Especially in interactive context. That can cause some permission errors:

In [1]: pd.Series([1,2,3]).to_hdf('foo.h5', 'x')

In [2]: pd.Series([1,2,3]).to_hdf('foo.h5', 'y')

In [3]: pd.HDFStore('foo.h5', 'r').keys()
Out[3]: ['/x', '/y']

In [4]: pd.Series([1,2,3]).to_hdf('foo.h5', 'z')
ValueError: The file 'foo.h5' is already opened, but in read-only mode.  Please close it before reopening in append mode.

In [5]: os.rename('foo.h5', 'bar.h5')
On Windows:
PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'foo.h5' -> 'bar.h5'

Feature Description

Adding destructor solves this

pd.io.pytables.HDFStore.__del__ = lambda self: self.close()

Alternative Solutions

.

Additional Context

No response

Comment From: snitish

Context managers and explicitly calling .close() are the two recommended ways to deal with objects like HDFStore. I don't think it is good practice to rely on the __del__ method, as it is not guaranteed to be called.

Comment From: mroeschke

Agreed with @snitish. Since HDFStore already has APIs for closing, closing this issue