piccolo-orm
Column types unknown to Piccolo crash schema generation in default look-up.
Piccolo doesn't handle the (fixed length) Character datatype.
When the column type is unknown to Piccolo, the schema generation defaults to the generic Column type, and a warning is added to the output.
Later on the generator attempts to parse the default value for the column. Unfortunately the COLUMN_DEFAULT_PARSER map does not contain an entry for Column, causing the generation to fail with a key-error.
Solution is to use a get(). Near line 358: pat = COLUMN_DEFAULT_PARSER.get(column_type)
Same is true for the Array datatype which is also currently not handled by Piccolo.
@gmos You're right - good catch.
I'm surprised the unit tests didn't catch this, I need to investigate.
I've opened a PR which makes the change you suggested.
It would be fairly simple to add a Character column type to Piccolo. I don't tend to use them - what kinds of things do you usually use them for?
@dantownsend I use Character type for fixed length ID's of sensors. Saves one (or two) length bytes per item. Is a wee bit more economical on storage, memory and CPU and comparisons, sorts etc will be bit faster. Actually irrelevant for my application with only 1000 or so sensors. More a habit from long ago. I'll happily change them to varchar in the DB.
But the absence of Array can't be worked around. Not having a crash anymore is ok for now. Support for arrays would be nice, but I have it only once.