Better handling of Target Schema Changes

Been a while - things are just working!

I am hoping at some point in time that there can be a discussion - enhancements regarding schema changes. Currently, the dataflow stops with an error that the schema changed. This is of no help in determining what changed.

I am assuming that internally the tool is looking at field positions and not the field name. I am also assuming that using field positions is used to increase performance? If it were based on the field name then changes to the schema that resulted in no change to the field name would work.

It would be nice to at a minimum be able to see what has changed and how it affects the order. Ideally a warning of a different schema and successful completion of the dataflow would be ideal.

This issue is costing me enough time that I am ending up redesigning the whole project to work with intermediary tables so that new table schemas will not consume to much time to resolve as I can remap from the intermediary tables to the new schema without having to do validation and transformations.

We can discuss it right now but I do not understand your proposal...

The simple solution is to put another transformer and use auto-map function.

Mike

Latest version provides better messages for the users

also

V 5.7.4.0 changes:

Massive performance boost

+ Text to INTERBASE TABLE up to 25.50%
+ Text to MS SQL TABLE via BCP up to 32.64%
+ Text to MS SQL TABLE via OLE DB 2000 CLIENT up to 50.00%
+ Text to MS SQL TABLE via OLE DB 2005 CLIENT up to 50.00%
+ Text to MS SQL TABLE via OLE DB 2008 CLIENT up to 50.00%
+ Text to MS SQL TABLE via OLE DB 2012 CLIENT up to 50.00%
+ Text to MY SQL TABLE up to 20.61%
+ Text to ORACLE TABLE via Conventional Path Loading up to 24.40%
+ Text to ORACLE TABLE via Direct Path Loading up to 30.45%
+ Text to ORACLE TABLE via OLE DB up to 19.52%
+ Text to ORACLE TABLE via OLE DB using ODBC up to 30.00%
+ Text to POSTGRESQL TABLE up to 29.15%

+ Better handling of Target Schema Changes

+ It is now possible to use a package variable in web service url
+ It is now possible to use a package variable in defaults
+ Notes have pop up menu now

- Toolbox in Packages is movable and sizeable
- SQL Statements: Leading comments -- or /**/ cause SQL to error - corrected
- Transforms Field List: Ctrl-X, C, V, <Delete> Do not work (but right click does) - corrected
- Transforms File Names: Ctrl-X, C, V, <Delete> Do not work (but right click does) - corrected
- Transforms Data List: Can not Ctrl-C or Right Click Copy - corrected
- Log Viewer: Can not Ctrl-C or Right Click Copy - corrected
- Clarify day label on File Operations - corrected

Just got burned again by source data in a different order. Took a while to track down where the change was. Adding transformation did not resolve. I had to reorder the underlying table to get the mappings aligned back up.

I really do not understand why the app cannot look at the field names - which rarely if ever change and not the order which can change often. What is the technical reasoning for using the field order instead of the field name for the mappings?

One thing I think I just discovered is that you have to open and close EVERY item on the canvas for everything to fully update. This is so frustrating when simply using the field name instead of the field position would avoid the entire issue.

I really do not understand why the app cannot look at the field names - which rarely if ever change and not the order which can change often. What is the technical reasoning for using the field order instead of the field name for the mappings?

It is not just that simple...

Some of our financial customers wanted to have duplicated fields names.
So were were forced to use fields positions.

Mike