Error when read_json created by to_json with orient='table and index=False

[albertvillanova]

Code Sample, a copy-pastable example if possible

import pandas as pd
df = pd.DataFrame({'a': [1, 2], 'b': [3, 4]})
dfjson = df.to_json(orient='table', index=False)
result = pd.read_json(dfjson, orient='table')

Problem description

An error is raised while the original DataFrame should be recovered:

pandas/io/json/table_schema.py in parse_table_schema(json, precise_float)
    315     df = df.astype(dtypes)
    316 
--> 317     df = df.set_index(table['schema']['primaryKey'])
    318     if len(df.index.names) == 1:
    319         if df.index.name == 'index':

KeyError: 'primaryKey'

My pandas version is 0.24.0.

Expected Output

The expected output result should be the original DataFrame df.

Output of pd.show_versions()

pandas/util/_print_versions.py in (mod) 78 ("blosc", lambda mod: mod.__version__), 79 ("bottleneck", lambda mod: mod.__version__), ---> 80 ("tables", lambda mod: mod.__version__), 81 ("numexpr", lambda mod: mod.__version__), 82 ("feather", lambda mod: mod.__version__),

AttributeError: module 'tables' has no attribute 'version'

[WillAyd]

Thanks for the report. Investigation and PRs are always welcome.