Hi, I have a case where the source table has a 

JSON

 column and Meltano unpacks it into multiple columns, not even loading the source column itself. I tried getting it by only specifying that single column to load, ran 

meltano elt

 and I cannot seem to be able to load that column into my schema. I only get the unpacked version (meaning all the columns that derive from that, according to Meltano). Example:

source: my_source_table

specification in meltano.yml:
select:
-source.specific_json_col

run command: 
meltano elt tap-bla target-bla-bla

what I see when loaded:

my_source_table:
specific_json_col__derived_col_1
specifi_json_col__derived_col_2
..

I did not look at the source code yet, but is this how it is meant to work? 🤔 Thanks in advance 🙂

Hi @rigerta! To answer your question, yes this it is normally meant to work simple smile. If you want the target not to unpack the json object into multiple columns, you'd have to remove any

properties

from the schema of the field in the source. Like

"specific_json_col": {
  "type": "object",
  "properties": {
    "derived_col_1": {...},
    "derived_col_2": {...}
  }
}

into

"specific_json_col": {
  "type": "object",
  "properties": {}
}

Depending on the target, it may have a setting option to disable flattening of json fields

I see, thanks @edgar_ramirez_mondragon!

np!