python 2 创建的 h5 文件数据到 python 3

Question

I am in the progress of transferring some piece of code from python 2 (2.7) to python 3 (3.7 or later)

However this piece of code reads a h5 file which was created by code in python 2.7. This piece of code will also be transferred to python 3, but not by me. I need the data in the h5 file to check whether the conversion to python 3 on my end works well (internally the data is a pandas dataframe).

Therefore I am looking for a trick (using either python 2 or python 3) to convert this h5 file into something that I can than read with python 3. It does not need to be a neat solution since it will only be temporarily.

The data is rather sizable.

Answer 1

So what I ended up doing is using python 2 to read the h5 and storing it as a json (one per key in the h5)

Then I used a python 3 script to read the jsons and store them as an h5 file again

(in python 2)
foo = pandas.read_hdf('file.h5', key='bla', mode='r')
foo.to_json('file.json')

(in python 3)
foo = pandas.read_json('file.json')
foo.to_hdf('file2.h5', key='bla', mode='w')

So it ended up being rahter simple. Hopefully this asnwer will help someone being stuck with the same.