Druid data loader does not work with input time column name - '__time'

[AshishKapoor]

Druid data loader does not work with input time column name - '__time'

Affected Version

The Druid version where the problem was encountered.
v0.19.0

Description

Please include as much detailed information about the problem as possible.

• Cluster size: NA

• Configurations in use paste data

• Steps to reproduce the problem
  Use this dataset, go to druid console, load data -> paste data -> paste following data-

{ "__time": "2013-08-31T01:02:33Z", "page": "test" }

• The error message or stack traces encountered. Providing more context, such as nearby log messages or even entire logs, can be helpful.

unparseable
Found unparseable columns in row: [org.apache.druid.indexing.overlord.sampler.SamplerInputRow@487bb21f], exceptions: [could not convert value [2013-08-31T01:02:33Z] to long,]
Original row: { "__time": "2013-08-31T01:02:33Z", "page": "test" }

Screenshot

• Any debugging that you have already done
  NA

Sincerely
Ashish Kapoor

[FrankChen021]

__time is a reserved name representing time column of the segment. Currently it requires a value of of unix epoch. You could change the property name of __time in json to another name.

[AshishKapoor]

__time is a reserved name representing time column of the segment. Currently it requires a value of of unix epoch. You could change the property name of __time in json to another name.

@FrankChen021 I've noticed it works with other databases.