Tags in write_points: tag_columns not optional when using json protocol

[eduisterwinkel]

• InfluxDB-python version: 5.3.1
• Python version: tried 3.6 and 3.8
• Operating system version: Windows 10

DataFrameClient.write_points using protocol=json currently only seem to embed tags when also tag_columns are specified. This is not what you would want, and also not as stated in the docs.

influxdb-python/influxdb/_dataframe_client.py

Lines 286 to 313 in 7cb5656

	if not tag_columns:
	points = [
	{'measurement': measurement,
	'fields':
	rec.replace([np.inf, -np.inf], np.nan).dropna().to_dict(),
	'time': np.int64(ts.value / precision_factor)}
	for ts, (_, rec) in zip(
	dataframe.index,
	dataframe[field_columns].iterrows()
	)
	]
	return points
	points = [
	{'measurement': measurement,
	'tags': dict(list(tag.items()) + list(tags.items())),
	'fields':
	rec.replace([np.inf, -np.inf], np.nan).dropna().to_dict(),
	'time': np.int64(ts.value / precision_factor)}
	for ts, tag, (_, rec) in zip(
	dataframe.index,
	dataframe[tag_columns].to_dict('record'),
	dataframe[field_columns].iterrows()
	)
	]
	return points

Please make tag_columns also optional using json protocol.

[eduisterwinkel]

Furthermore,

influxdb-python/influxdb/_dataframe_client.py

Line 308 in 7cb5656

dataframe[tag_columns].to_dict('record'),

needs to be:
dataframe[tag_columns].to_dict(orient='records'),
with the 1.3.4 pandas version I am using (not sure since which version of pandas)