different results for tUniqueRow and tAggregateRow

[Anonymous]

Hi guys,

I have dataset with only one column and number of rows. Why I'm getting different result for tUniqueRow and tAggregateRow ?

Thanks !

[vapukov]

may be because - one is count Unique rows and other any rows? 🙂

in other words - You have duplicates in this filtered column

[Anonymous]

But the input to both the components is same and mainly has only one column, and yes this one column is having duplicate values in it. How can the result be different ?

I think tUniqueRow is using fuzzy match for uniqueness, and so the no. of rows are lesser by 33 than that of tAggregateRow.

Thanks for the reply @vapukov.

[vapukov]

Unique - it is Unique 🙂

1,1,2,3,3,3 = unique 1,2,3 without variants

[Anonymous]

Okay 

You mean to say for 1,2,3,3,4,4,5,5,5,6,6,6,6

Unique = 1,2

Aggregate = 1,2,3,4,5,6 

Right ? @vapukov

[vapukov]

no, 

I guess, what are You mean

if You have only single column and group by this column, it must be same result
because not - need redirect flows to files and compare by diff

for make it more easy compare - You can sort column before store to file

what settings You use in both components?

[cterenzi]

What is the data type of the single column in your schema?  How do you have your tAggregateRow component configured?

[Anonymous]

@vapukov, below are the screenshot of basic setting of both the componnets

tUniqueRow setting:

tAggregateRow setting:

@cterenzi, data dype of single column 'VENDOR_NAME' is String.

[cterenzi · Accepted solution]

I was hoping you'd say it was BigDecimal or something and this was a precision issue.

tUniqRow by default does a case insensitive comparison. If your data has different casing for the same vendor name, this would result in fewer values output by tUniqRow. If you want tUniqRow to be case sensitive, you can check the box in its settings (next to Key attribute).

[Anonymous]

Oh, I haven't observed that. Thanks for pointing me out @cterenzi.

Now I'm getting same row count. Which one is good to go for - tUniqueRow or tAggregateRow ?