> PS I tried the 'Add restriction to table' and it's just wicked, > extremely useful! My particular use case is that I have an Experiment > table with different types of experiment-thingies: frequency > determination, association analysis etc. This function let's me > easily create seperate datasets for each!
I just saw this... sadly, this is gone in 0.6! But all is not lost as it has been superceded by 'Add restriction to relation' instead. I figured that having them both was duplication of effort, quite apart from the nightmare that 'add restriction to table' presented at the SQL transformation stage... :) The 'add restriction to relation' works very similarly. It is applied to a relation (or a particular iteration of that relation if the relation is compounded, see previous email...), and involves choosing columns from one or both ends of the relation to include in some kind of expression. (You don't have to include both ends - so in your case you would only include columns from your Experiment table.) These are in addition to the columns in the keys that the relation joins, which cannot be changed. You can then tick a box labelled 'hard restriction' which when left unticked makes the relation a left-join (by default all relations in MBuilder are made into left-joins), and when ticked makes it an inner-join, therefore excluding (or nulling-out) entire rows from the transformed table that don't match the conditions. 'Concat relation' has also gone as it just didn't work properly in any but the simplest cases, and required highly database-specific code to work which was a nightmare to test and maintain. Other new things for 0.6 are a much improved and fully interactive 'explain table' and 'explain dataset' dialog, corrected and extended optimiser (boolean/count) column behaviour, and the introduction of MartRunner, a tool which will take the generated SQL and run it for you, giving you progress reports and timings as it goes. Enjoy! cheers Richard
