BIML GetDropAndCreateDdl creating incorrect lengths for datatypes - tsql

Question

When I use GetDropAndCreateDdl to generate CREATE scripts for tables, I get datatypes for columns that are different than what the datatypes actually are.
This results in a package validation error that "the error output has properties that do not match the properties of its corresponding data source column" and a validation status of "VS_NEEDSNEWMETADATA".
If I right-click the connection source, select Show Advanced Editor, and take a look at the Column Mappings, I can see that [Column1] in the Available External Columns list has a different Length than the datatype that was generated in the GetDropAndCreateDdl. I can delete and re-create the metadata mappings, but that is not a viable solution since there are many dataflow tasks.
How do I get GetDropAndCreateDdl to create the correct datatypes with correct lengths?
I am using ImportDB to get the list of tables, metadata, etc.
Environment.biml
<Biml xmlns="http://schemas.varigence.com/biml.xsd">
<Connections>
<OdbcConnection Name="OdbcSrc_DV" ConnectionString="Dsn=Source-32bit-test;" />
<OleDbConnection Name="OleDbDst_Staging" ConnectionString="Provider=SQLNCLI11;Server=SQL-DEV;Initial Catalog=Source_Staging;Integrated Security=SSPI;" />
</Connections>
<Databases>
<Database Name="Source" ConnectionName="OdbcSrc_DV" />
<Database Name="Source_Staging" ConnectionName="OleDbDst_Staging" />
</Databases>
<Schemas>
<Schema Name="dbo" DatabaseName="Source" />
<Schema Name="dbo" DatabaseName="Source_Staging" />
</Schemas>
</Biml>
CreateTableMetadata.biml
<## import namespace="System.Data" #>
<## import namespace="Varigence.Biml.CoreLowerer.SchemaManagement" #>
<#
var sourceConnection = RootNode.DbConnections["OdbcSrc_DV"];
var importResult = sourceConnection.ImportDB("", "", ImportOptions.ExcludeForeignKey | ImportOptions.ExcludeColumnDefault | ImportOptions.ExcludeViews);
var tableNamesToImport = new List<string>() { "Test_Table" };
#>
<Biml xmlns="http://schemas.varigence.com/biml.xsd">
<Tables>
<# foreach (var table in importResult.TableNodes.Where(item => tableNamesToImport.Contains(item.Name)).OrderBy(item => item.Name)) { #>
<Table Name="<#=table.Name#>" SchemaName="Source.dbo">
<Columns>
<#=table.Columns.GetBiml()#>
</Columns>
<Annotations>
<Annotation AnnotationType="Tag" Tag="SourceSchemaQualifiedName"><#=table.SchemaQualifiedName#></Annotation>
</Annotations>
</Table>
<# } #>
</Tables>
</Biml>
DeployTargetTables.biml
<## template tier="2" #>
<Biml xmlns="http://schemas.varigence.com/biml.xsd">
<Packages>
<Package Name="MasterTableDeploy" ConstraintMode="Parallel">
<Tasks>
<# foreach (var table in RootNode.Tables) { #>
<ExecuteSQL Name="SQL CREATE <#=table.Name#>" ConnectionName="OleDbDst_Staging">
<DirectInput><#=table.GetDropAndCreateDdl()#></DirectInput>
</ExecuteSQL>
<# } #>
</Tasks>
</Package>
</Packages>
</Biml>
CreateLoadPackages.biml
<## template tier="2" #>
<Biml xmlns="http://schemas.varigence.com/biml.xsd">
<Packages>
<Package Name="Copy Data" ConstraintMode="Parallel">
<Tasks>
<# foreach (var table in RootNode.Tables) { #>
<ExecuteSQL Name="SQL TRUNCATE <#=table.Name#>" ConnectionName="OleDbDst_Staging">
<DirectInput>TRUNCATE TABLE <#=table.Name#></DirectInput>
</ExecuteSQL>
<Dataflow Name="DFT LOAD <#=table.Schema.Name#>_<#=table.Name#>">
<PrecedenceConstraints>
<Inputs>
<Input OutputPathName="SQL TRUNCATE <#=table.Name#>.Output" />
</Inputs>
</PrecedenceConstraints>
<Transformations>
<OdbcSource Name="ODBC_SRC <#=table.Name#>" Connection="OdbcSrc_DV">
<DirectInput>SELECT <#=table.GetColumnList()#> FROM <#=table.GetTag("SourceSchemaQualifiedName")#></DirectInput>
</OdbcSource>
<OleDbDestination Name="ODBC_DST <#=table.Name#>" ConnectionName="OleDbDst_Staging">
<TableOutput TableName="<#=table.ScopedName#>" />
</OleDbDestination>
</Transformations>
</Dataflow>
<# } #>
</Tasks>
</Package>
</Packages>
</Biml>
Here is the script that gets created from the ODBC source -
And here are the results from INFORMATION_SCHEMA.COLUMNS for the same table -

Answer 1

Have you tried using the newer method: GetDatabaseSchema? I've found it to be much more reliable across different connection types.
http://www.cathrinewilhelmsen.net/2015/07/12/biml-extension-methods-getdatabaseschema/