On 4/17/18, petern wrote:
> 3. In tables with serially related data it can be cumbersome to reference
> columns in the working table when columns must be computed from existing
> rows in the same table.
I don't understand this point. Can you provide an example?
--
Some points on working table alias:
1. If upsert goes to trunk, there is already a INSERT table alias test case
(do_execsql_test upsert3-210) that must succeed:
https://www.sqlite.org/src/info/907b5a37c539ea67
2. Obviously PostgreSQL already supports a working table alias universally:
Hello Richard !
Now that you are making changes on sqlite3 parser could you please add
the table alias to delete/insert/update ?
I already have it working and got conflicts with the latest changes,
they are not big or complicated but it helps make some queries easier.
Cheers !
fossil diff
3 matches
Mail list logo