You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Loading json straight into ES is part of the road-map, probably for M2. There are several performance considerations here since loading the file through Hive or Pig would only slow things down and also confuse the translation (json itself is not recognized but rather translated as strings which is not what one wants).
One could use the Pig json integration to load the file but that would only complicate the workflow w/o any benefits (quite the opposite).
I'm currently thinking of bypassing all layers and reading the resource directly and pass it to ES.
I see how we can index a flat structured file, but is there a way to index if file already has data in a json format? For eg:
{
user {
phones { [1,2,4]
}
}
}
If this is not possible then I can probably contribute and add this functionality.
One way I can think of is to read each record as a charray in pig and directly index it like: Not sure if this will work though
A= LOAD 'file' using PigStorage (data chararray);
Store A into ...
The text was updated successfully, but these errors were encountered: