Alex Amato created BEAM-618: ------------------------------- Summary: Python SDKs writes non RFC compliant JSON files for BQ Export Key: BEAM-618 URL: https://issues.apache.org/jira/browse/BEAM-618 Project: Beam Issue Type: Bug Components: sdk-py Reporter: Alex Amato Assignee: Frances Perry
Python SDK uses the built in json.dumps to write JSON files to GCS for the BQ Exporter. BigQuery can fail to parse these files when it tries to load these files into a BQ table because json.dumps can export JSON which does not conform to the IEEE RFC. There are a few cases which are not RFC compilant listed in that module. https://docs.python.org/2/library/json.html#standard-compliance-and-interoperability The main issue we run into is the NAN, INF and -INF values. These fails with a confusing error (and we delete the GCS files making it hard to debug): JSON table encountered too many errors, giving up. Rows JSON parsing error in row starting at position We can set the allow_nan argument to json.dumps to false to address these issues. So that when a user tries to write a file with INF, -INF or NAN Setting this argument will produce this type of error when json.dumps is called with NAN/INF values. We may want to catch this error to mention the fact that INF and NAN are not allowed. Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.7/json/__init__.py", line 250, in dumps sort_keys=sort_keys, **kw).encode(obj) File "/usr/lib/python2.7/json/encoder.py", line 207, in encode chunks = self.iterencode(o, _one_shot=True) File "/usr/lib/python2.7/json/encoder.py", line 270, in iterencode return _iterencode(o, 0) ValueError: Out of range float values are not JSON compliant -- This message was sent by Atlassian JIRA (v6.3.4#6332)