Copy-DbaDbTableData -Query still requires -Table, with an unhelpful error otherwise

[JankeUwe]

Environment: dbatools 2.8.2 (also seen on 2.7.1), SQL Server

Description:
Copy-DbaDbTableData in -Query mode still requires -Table to be specified, even though the query itself already determines what's being read from the source. Omitting -Table when using -Query produces a generic error that doesn't mention -Query at all:

You must pipe in a table or specify SqlInstance, Database and [View|Table]

Repro:

• Run Copy-DbaDbTableData -SqlInstance <src> -Database <db> -Query "SELECT ..." -DestinationSqlInstance <dst> -DestinationDatabase <db> -DestinationTable <table> without -Table.
• Result: the error above, which reads as if -SqlInstance/-Database were missing (they weren't) rather than pointing at the actual missing parameter.

Where this bit us:
Hit this while chunked-transferring a large partitioned table (date-column chunking via -Query), the most expensive case to get an unclear error on.

Expected:
Either drop the -Table requirement when -Query is supplied (since the query already defines the source), or make the error message explicitly say that -Table is required alongside -Query in this mode.

Workaround used:
Always pass -Table alongside -Query, even though the query is what actually determines the source data.

[andreasjordan]

Confirmed on development and fixed the error-message half in the PR above: with -Query and no -Table/-View the command now says explicitly that the table or view is still required, instead of a message that reads as if -SqlInstance or -Database were missing.

On dropping the requirement: the table or view is doing more work in -Query mode than it looks like. It supplies the source connection context, the structure that -AutoCreateTable scripts, the default destination table name, and the source columns of the output object. Removing it means answering what all of those should be when only a query exists - that is a feature discussion rather than a bug fix, so the requirement stays for now and the help for -Query continues to document it.

For your chunking scenario the practical shape is: pass the partitioned table itself as -Table alongside each chunk's -Query - it is the correct metadata source anyway, and with #10666 merged the query columns then map positionally onto the writable destination columns.

created by Claude and reviewed by Andreas Jordan