sqlpackage.exe SELECT statement causing massive reads

We are using sqlpackage.exe to deploy scripts onto a SQL Server 2016 SP1 instance. Running sqlpackage.exe executes this SELECT

```
SELECT * FROM (
SELECT
SCHEMA_NAME([o].[schema_id]) AS [SchemaName],
[si].[object_id] AS [ColumnSourceId],
[o].[name] AS [ColumnSourceName],
[o].[type] AS [ColumnSourceType],
[ic].[column_id] AS [ColumnId],
[c].[name] AS [ColumnName],
[si].[index_id] AS [IndexId],
[si].[name] AS [IndexName],
[ds].[type] AS [DataspaceType],
[ds].[data_space_id] AS [DataspaceId],
[ds].[name] AS [DataspaceName],
[si].[fill_factor] AS [FillFactor],
[si].[is_padded] AS [IsPadded],
[si].[is_disabled] AS [IsDisabled],
[si].[allow_page_locks] AS [DoAllowPageLocks],
[si].[allow_row_locks] AS [DoAllowRowLocks],
[sit].[cells_per_object] AS [CellsPerObject],
[sit].[bounding_box_xmin] AS [XMin],
[sit].[bounding_box_xmax] AS [XMax],
[sit].[bounding_box_ymin] AS [YMin],
[sit].[bounding_box_ymax] AS [YMax],
[sit].[level_1_grid] AS [Level1Grid],
[sit].[level_2_grid] AS [Level2Grid],
[sit].[level_3_grid] AS [Level3Grid],
[sit].[level_4_grid] AS [Level4Grid],
[sit].[tessellation_scheme] AS [TessellationScheme],
[s].[no_recompute] AS [NoRecomputeStatistics],
[p].[data_compression] AS [DataCompressionId],
CONVERT(bit, CASE WHEN [ti].[data_space_id] = [ds].[data_space_id] THEN 1 ELSE 0 END)
AS [EqualsParentDataSpace]
FROM
[sys].[spatial_indexes] AS [si] WITH (NOLOCK)
INNER JOIN [sys].[objects] AS [o] WITH (NOLOCK) ON [si].[object_id] = [o].[object_id]
INNER JOIN [sys].[spatial_index_tessellations] [sit] WITH (NOLOCK) ON [si].[object_id] = [sit].[object_id] AND [si].[index_id] = [sit].[index_id]
INNER JOIN [sys].[data_spaces] AS [ds] WITH (NOLOCK) ON [ds].[data_space_id] = [si].[data_space_id]
INNER JOIN [sys].[index_columns] AS [ic] WITH (NOLOCK) ON [si].[object_id] = [ic].[object_id] AND [si].[index_id] = [ic].[index_id]
INNER JOIN [sys].[columns] AS [c] WITH (NOLOCK) ON [si].[object_id] = [c].[object_id] AND [ic].[column_id] = [c].[column_id]
INNER JOIN [sys].[objects]

Disabling Query Optimizer Hot Fixes resolves the problem here.

This means I'm using the 2016 cardinality estimator minus any improvements to the Query Optimizer brought on by other service packs and hotfixes?

I never quite understood what happens when you disable hotfixes in the 130 compatibility level.

Here's the explanation from the KB. Just adds to the confusion in my opinion.


Controls query optimizer changes released in SQL Server Cumulative
Updates and Service Packs. Starting with SQL Server 2016, trace flag
4199 changes that are made to previous releases of SQL Server will
become enabled under database compatibility level 130 without trace
flag 4199 enabled. For more information, see this Microsoft Support
article.

Which then leads to this article: sql-server-query-optimizer-hotfix-trace-flag-4199-servicing-model

They suggest option three which is what I am now using.

StackExchange Database Administrators Q#163703, answer score: 2