Hi, I have a question about overriding data types. Is it possible to set the data type only with the environment variable without first using

meltano config <extractor> set _schema <entity> <attribute> <key> <value>

command? My env var

TAP_SPREADSHEETS_ANYWHERE__SCHEMA_LIST_OF_PRODUCTS_PRODUCT_NUMBER_TYPE=["string", "null"]

is only considered if I have the following schema section in `meltano.yml`:

plugins:
  extractors:
  - name: tap-spreadsheets-anywhere
    namespace: tap_spreadsheets_anywhere
    pip_url: git+<https://github.com/ets/tap-spreadsheets-anywhere.git>
    capabilities:
    - catalog
    - discover
    - state
    schema:
      list_of_products:
        product_number:
          type:
          - string
          - 'null'

@jan_fertek What you've found is correct: it's not currently possible to set a setting using an env var until Meltano knows about that particular setting because of it being defined in

meltano.yml

. This is because Meltano only defines the top-level

_schema

extra (with

<TAP>__SCHEMA

), and doesn't know about any of the nested settings until they're defined as custom settings and get an env var: https://meltano.com/docs/configuration.html#custom-settings Meltano doesn't currently parse the "original" schema from the catalog to proactively detect the nested settings and determine their env vars, nor does it eagerly read from the environment and attempt to parse

TAP_SPREADSHEETS_ANYWHERE__SCHEMA_LIST_OF_PRODUCTS_PRODUCT_NUMBER_TYPE

into

_schema.list_of_products.product_number.type

, since it's impossible to know just from that string which underscores represent nesting, and which represent actual underscores.