![]() ![]() This means that the OLEDB data source itself has failed to change its output type. I'm doing this in SSIS 2012:ġ) Double-click the "Data Flow Path" error coming out of the OLEDB source, look at the "Metadata", and verify that the type of your column(s) is the wrong, old data type. I can give you one solution to this, in the case where your data source is an OLEDB data source and the database entity driving this source changed its emitted data type (but the OLEDB data source is still emitting the old type). There should be a checkbox about dropping unmatched columns or something like that. Make sure your dataflow is fully connnected from source to destination, double click on an arrow and map the columns. So that's the problem you are seeing: it remembers columns it shouldn't remember because the source has changed. It now "remembers" metadata, so you can build your dataflow from destination to source, instead of the usual way around. They redesigned SSIS in SQL Server 2012 and the way it handles metadata in the dataflow. I have tried a bunch of different fake outs only thing that works is dropping the OLE DB Destination and adding a new one. Refresh metadata via pointing to a table/statement with no common columns still seems to work for source, columns are reset and ordered as expected.ĭestination updated but still retains ALL columns from Customer table, regardless of pointing to new destination with no common columns and then pointing back. OLE DB Source is updated to point to Account table. Package is copied and pasted, guid's reset, renamed to Package B. Package A loads Customer data with OLE DB Source and OLE DB Destination. ![]() It renders leveraging the copy & pasting of source and destination components fairly useless. ![]() ![]() I am curious if anyone else is seeing this, and more curious if there is a workaround. So in SSIS 2012 (VS2010 using SSDT) I am noticing an issue with metadata refresh that no longer responds to the tricks described in this post, and worse still, the metadata retains uncommon columns, as well. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |