PyTables: BUG: setting a numpy array writeable not supported in numpy>=1.16

In reading hdf5 data, this line is setting the writable flag of an ndarray to True which is dangerous and no longer allowed as of numpy 1.16. Is the line needed? The buf is a new copy, so it should be readwrite.

xref numpy/numpy#12791

About this issue

Original URL
State: closed
Created 5 years ago
Comments: 15 (8 by maintainers)

Links to this issue

pandas - Can't load HDF5 files in Python - Stack Overflow

Commits related to this issue

Adjust required package versions * 2 missing packages: xlrd, Rtree * pandas: needs to be >0.21.0 as the column param "columns" in drop() which is used in reegis is not supported in versions<0.21.0 ... — committed to nesnoj/reegis by nesnoj 5 years ago
env: Update pytables to v2.5.1 There is an issue with numpy and pytables support which results in a problem reading files (https://github.com/PyTables/PyTables/issues/719). This updates the version o... — committed to malramsay64/statdyn-analysis by malramsay64 5 years ago

Most upvoted comments

Yeah, thanks for the reminder. Yes, we need to do a release as soon as possible, but I am currently swamped. Any volunteers?

FrancescAlted on Jan 18, 2019

The new 3.5.1 release is out and this should be fixed.

FrancescAlted on Mar 14, 2019

A new release would be really nice…

Thanks!

stevenwinfield on Feb 14, 2019

@FrancescAlted it seems to me that this issue is a duplicate of #718 and #717. It has been already fixed in f6b64c9.

Probably the current fix can be improved but it already works AFAIK.

Probably we need to make a new delivery.

I’m closing. Please feel free to reopen if you think there is still some open point.

avalentino on Jan 18, 2019