BTW this would be cool to do as well: ALTER TABLE ks.to_copy LIKE ks.tb WITH INDICES;
This would mean that if we create a copy of a table, later we can decide that we need indices too, so we might "enrich" that table with indices from the old one without necessarily explicitly re-creating them on that new table. On Mon, Aug 19, 2024 at 6:55 PM Štefan Miklošovič <smikloso...@apache.org> wrote: > I think this is an interesting idea worth exploring. I definitely agree > with Benjamin who raised important questions which needs to be answered > first. Also, what about triggers? > > It might be rather "easy" to come up with something simple but it should > be a comprehensive solution with predictable behavior we all agree on. > > If a keyspace of a new table does not exist we would need to create that > one too before. For the simplicity, I would just make it a must to create > it on same keyspace. We might iterate on that in the future. > > UDTs are created per keyspace so there is nothing to re-create. We just > need to reference it from a new table, right? > > Indexes and MVs are interesting but in theory they might be re-created too. > > Would it be appropriate to use something like this? > > CREATE TABLE ks.tb_copy LIKE ks.tb WITH INDEXES AND VIEWS AND TRIGGERS > .... > > Without "WITH" it would just copy a table with nothing else. > > On Mon, Aug 19, 2024 at 6:10 PM guo Maxwell <cclive1...@gmail.com> wrote: > >> Hello， everyone： >> As Jira CASSANDRA-7662 >> <https://issues.apache.org/jira/browse/CASSANDRA-7662> has described , >> we would like to introduce a new grammer " CREATE TABLE LIKE " >> ,which simplifies creating new tables duplicating the existing ones . >> The format may be like : CREATE TABLE <new_table> LIKE <old_table> >> >> Before I implement this function, do you have any suggestions on this? >> >> Looking forward to your reply！ >> >
