Partial support for virtual columns #6919
Conversation
…n converting virtual columns on store / read
|
PTAL @reltuk |
| { | ||
| Name: "merge a virtual column", | ||
| SetUpScript: []string{ | ||
| "create table t1 (id bigint primary key, v1 bigint, v2 bigint, v3 bigint as (v1 + v2), index (v3))", |
There was a problem hiding this comment.
was there a different PR for index support? do we have standalone tests for that?
There was a problem hiding this comment.
You're not missing anything, index merges are broken. They're non trivial to fix because we need the engine's help to rebuild them. I just want to get this in before it gets any bigger, going to do another pass to fix indexes / merge and add more tests.
| }, | ||
| }, | ||
| { | ||
| Name: "merge a virtual column created on another branch", |
There was a problem hiding this comment.
what happens if you have a schema merge conflict between virtual/generated/real column that are otherwise the identical in type/name/value? do you have to pick one?
There was a problem hiding this comment.
Good question for the next PR
| @@ -864,6 +864,8 @@ func getTableSchemaAtRef(queryist cli.Queryist, sqlCtx *sql.Context, tableName s | |||
| return sch, createStmt, nil | |||
| } | |||
|
|
|||
| // schemaFromCreateTableStmt returns a schema for the CREATE TABLE statement given | |||
| // TODO: this is substantially incorrect, doesn't handle primary key ordering, probably other things too | |||
There was a problem hiding this comment.
This probably bothers me the most in this PR. Is this func silently breaking things?
There was a problem hiding this comment.
Yes, we are for sure printing schema diffs wrong. Will add bats tests on my next pass
Merges don't work yet, coming in a separate PR.
But storage format details should be final.