just used this feature for the first time -- nice new feature to have, but I ran into a number or shortcomings; would be nice if these were addressed in a future release...
-- can't load into a view, or into a table owned by another schema (must spoof Apex by a creating table in workspace schema with the same structure as the view, create Data Load pages in Apex, then drop the table and create the view)
-- can't change the table name once a Data Load Definition is created (yet can redefine the PK -- doesn't make a lot of sense)
-- can't set any defaults for mapping fields to columns (only exact match is automatically mapped)
-- can't create a new Data Load Definition without rerunning the wizard (yet the Data Load Definition is treated as a shared object in the page process for the Source page and the region for the Mapping page, etc.)
-- page cache and collection are not cleared after the load is processed (however, the developer can modify the pages to manually fix this -- otherwise the Validation page can redisplayed and Load Data button can be selected again, resulting in duplicate data or insert errors)
thanks for your feedback about the new feature, I will briefly give my comments below
-- can't load into a view, or into a table owned by another schema
the data upload feature allows you to create data load defnition on all tables under the schemas that you are having access to. in our future release access to the views also will be possible.
-- can't change the table name once a Data Load Definition is created
The concept of data load definition itself is linked to table. table lookups and transformations rules are all defined towards a table name, hence they would not be any need to change the table name only.
You will most probably need to add/edit or remove table lookups and transformations rules Or re-create the pages without re-creating the data load definition. and you can do all the above through shared component.
-- can't set any defaults for mapping fields to columns
Usually, the mapping is left to the user, so that he can make sure he is doing the right matches between his data and the table column names but the data upload feature helps in matching the table columns with the names of columns that appear on user data.
-- can't create a new Data Load Definition without rerunning the wizard
I just explained above, Yes you can re-create the wizard without going through all the steps. Just go to the shared components and choose your data definition, you should see a link on the right corner - Re-create Data Load Pages
-- page cache and collection are not cleared after the load is processed
In future release, I beleive we will give an option to the developer to choose if he would like to maintain his collections or not. The idea of keeping the collections was to give the developer much more control of the data in case, he would like to do more or extend the feature.
The data loading wizard is a really good idea and just what we need to reduce our workload. I would like to request a few more features which would enable us to fully use it within our business.
1) when i upload a csv not all the columns display in the Data / Table Mapping screen. i assume there is a limit on the number of columns you can upload. is there any way you can highlight that this limit is exceeded when we select the Data Load Table we want to use? what is this limit?
2) can we have more control over the transformation lookup. ideally i would be able to write plsql expressions to run against each field during the upload.
3) can we have the ability to create multiple transformation runs in one go (i had 20 columns which needed the same transformation rule)
4) can the table lookup look at views? my table lookup needs to be a combination to two tables (test and live)