At dbdiagram.io you can specify a column without datatype like;
Table core.pk_tbl {
pk_col [pk]
}
This is accepted and “[pk]” is set as the datatype. I propose a column name followed by a bracket is not accepted by the parser but a “datatype is missing” error is raised.
If you add second [pk] clause it ids interpreted as the primary key notation.
Table core.pk_tbl {
pk_col [pk] [pk]
}
I think both are likely user errors and should not be accepted. Allowing malformed files on dbdiagram causes problems when you later try to use stricter parsers like pydbml.
edit: To allow brackets in datatype names, require it to be enclosed in quotes; “[pk]”.
Thank you for reporting, Jens
We have collected the issue and will try to improve it in the future.
Have a nice day!