Forum Discussion
smpa01
1 year agoCommunity Champion
Dynamically generate column schema as record
Is it possible to generate dynamically for all columns of a given table the following
[id=nullable Int64.Type, year=nullable Int64.Type, amount=nullable number, ing_est=nullable datetime] ...
- 1 year ago
Now I probably see what you want to achive. You want to replace this part of Table.Group. Is that correct?
If yes, you can just delete that part of Table.Group and use Table.Combine to preserve column types, like this:
let src=Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("lZBBCsMwDAT/4nMiVivLtvSO3EL+/4267aUlhWJdBIJhR3ueZUSUrRD0uXaFjjCprfX3te6IHfVAJDxJ0WaYU67tg60v1jkqhCugzYXlJOuNTuntm1RPjGQVRL+Tr//oGOai3WyF/aGpI9GTsytiRdOPmaNMhlhd1PzH3jT9UOQslE08FjVnzpN0ib6seWevBw==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [id = _t, year = _t, amount = _t, ing_est = _t]), ct = Table.TransformColumnTypes(src,{{"id", Int64.Type}, {"year", Int64.Type}, {"amount", type number}, {"ing_est", type datetime}}, "en-US"), GroupedRows = Table.Group(ct, {"id"}, {{"ad", each _, type table}}), Combined = Table.Combine(GroupedRows[ad]) in Combined
smpa01
1 year agoCommunity Champion
Smashing !!! Any performance implication you may think of for a large table?
dufoq3
1 year agoCommunity Champion
No, there shouldn't be any. Another advantage of Table.Combine is that it will expand each column of every table dynamicaly, besides if you Expand Columns via button - it will hardcode column names (but only for first table if I'm correct).