Dynamically generate column schema as record

[smpa01]

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]

sample query

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}}),
    #"Grouped Rows" = Table.Group(ct, {"id"}, {{"ad", each _, type table [id=nullable Int64.Type, year=nullable Int64.Type, amount=nullable number, ing_est=nullable datetime]}})
in
    #"Grouped Rows"

@AlexisOlson 

[dufoq3]

Hi @smpa01, check this:

Output:

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}, {"Schema", each Record.FromTable(Table.RenameColumns(Table.Schema(_)[[Name], [TypeName]], {{"TypeName", "Value"}}, MissingField.Ignore)), type record}})
in
    GroupedRows

[smpa01]

@dufoq3  this is good but I want to dynamically pass on that bit to finally give me the following

//didnot work
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 = let x = Table.Group(ct, {"id"}, {{"ad", each _, type table}, {"Schema", each Record.FromTable(Table.RenameColumns(Table.Schema(_)[[Name], [TypeName]], {{"TypeName", "Value"}}, MissingField.Ignore)), type record}})[Schema]{0} in Table.Group(ct, {"id"}, {{"ad", each _, type table x}})
in
    GroupedRows

[dufoq3]

What is the purpose of this? Do you want to replase table types with another table types or what? If yes, there is more efficient way:

let
    Source = #table(type table[Name=text, Age=Int64.Type, Date of Birth=date], {{"Peter", 20, #date(2000,10,15)}, {"Martin", 25, #date(2003,12,10)}}),
    ChangedType = Table.TransformColumnTypes(Source,{{"Age", type number}, {"Date of Birth", type datetime}}),
    RestoredTypes = Value.ReplaceType(ChangedType, Value.Type(Source))
in
    RestoredTypes

[smpa01]

The purpose is to create a function that can group any table with the dynamically generated schema of that table to be passed on to Table.Group.

[dufoq3 · Accepted solution]

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]

Smashing !!!  Any performance implication you may think of for a large table?

[dufoq3]

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).

[dufoq3]

Could you provide expected result based on sample data please?

[smpa01]

this is the end result I want to achieve once the schema is dynamically generated and passed on to Table.Group

The challenge is to dynamically somehow replace the following portion

[id=nullable Int64.Type, year=nullable Int64.Type, amount=nullable number, ing_est=nullable datetime]

 with a function call in the code below

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}}),
    #"Grouped Rows" = Table.Group(ct, {"id"}, {{"ad", each _, type table [id=nullable Int64.Type, year=nullable Int64.Type, amount=nullable number, ing_est=nullable datetime]}}),
    #"Expanded ad" = Table.ExpandTableColumn(#"Grouped Rows", "ad", {"id", "year", "amount", "ing_est"}, {"id.1", "year", "amount", "ing_est"})
in
    #"Expanded ad"

[PwerQueryKees]

Look at this

https://learn.microsoft.com/en-us/powerquery-m/table-schema