Data Type mappings - Picture with mappings in thre...
# singer-targets
v
Data Type mappings - Picture with mappings in thread. Writing the mssql target. The folks I'm working with want more specific mappings between data types in the source we are using (oracle) and target (mssql). From looking through a few other targets it seems like the singer approach is to take a generally correct view on the data types and move on? https://github.com/transferwise/pipelinewise-target-postgres/blob/master/target_postgres/db_sync.py#L40 While that seems appropriate for most integrations, and a needed default case, is there any spec out there on providing more specefic mappings? I'm thinking on top of (From a float in oracle)
Copy code
{
  "multipleOf": 1e-38,
  "type": [
    "null",
    "number"
  ]
}
A typehint added from the schema of oracle (catalog file oracle takes in has it as well)? like
Copy code
{"multipleOf": 1e-38, "type": ["null", "number"], sourceType:"oracle_FLOAT"}
My target would need specific mappings, and this is ugly but it works. Want to see if there's something out there first!
message has been deleted
Second column is MSSQL
Anyone have documentation on where
Copy code
"exclusiveMinimum": true, "exclusiveMaximum": true,
comes from in the schemas? Is this just made up per source? Looks like a convention from Singer? Is it documented?