JOIN Doubt

[ShantanuDagar]

If I use join tool on 2 datasets with 100 columns each and match every column from 1 dataset to corresponding another column in the configuration window. (All 100)

Will the output coming through J consist of dataset which is exactly same across every column in 2 datasets?

[binuacs]

@ShantanuDagar yes, the J-anchor will give you only those records which are matching in the join conditions. 

[IraWatt · Accepted solution]

@ShantanuDagar Yes that's right, the rows the come out of the J will be identical to one another. Though I would not recommend joining on 100 columns as that would take a while to do, I would recommend transposing your data first and joining on that instead. 

[MarqueeCrew · Accepted solution]

@ShantanuDagar ,

My recommendation is DON'T do that.  I would not create 100 join keys.  That's a bad idea.  There is a CReW macro for "Expect Equal" where if the data from the Expected Anchor equals the data from the Actual Anchor, then no error is created.

http://www.chaosreignswithin.com/p/macros.html

The data must be sorted in the same order for the comparison.  It will let you know if each record is identical.  If you JOIN the data and there are duplicate records, you could end up with a Cartesian join and create a huge output.  If there are empty rows, this could be worse.

If you must JOIN, then I'd create a single join key.

Cheers,

Mark