INTERVALMATCH Fails: Any other solutions?

[#1]

I attached a QVW that is giving me trouble.

I can run it for under 2000 TAP_EVENTS but if i try and run the thing my server cant handle it.

I am trying to break down the AUTHORIZED AMOUNT into daily amounts. The real trick is that each day has a specific value mon-fri are 1, sat and sun are .5 and holidays are 0 (So weeks are based on 6 day weeks, unless there is a holiday) i need each TAP_EVENT to be broken into a daily value.

The script I have is doing its job but QV cant handle more than the 2000 rows i have.

Any suggestions? Can QV handle this or should i use a other product for the "heavy lifting" (Business Objects or Informatica)?

Thanks.

[mike_garcia · Accepted solution]

You need to remove duplicate rows before doing the left join.

1) Do a Distinct Load of the DAY_DATE field BEFORE using it in the IntervalMatch Operation.
2) Also, do a Distinct Load of the two fileds making the interval while performing the Interval Match.

3) From the resulting table (after the IntervalMatch), do a Distinct load and perform the Left Join.

4) Drop the original table generated from the IntervalMatch (the first one from step #2)

[#3]

Hi, can anybody make this in example to. I'm struggeling with similar problem and can not do these distinct loads there.

LEFT

JOIN (FACTS)

INTERVALMATCH

( DATE, KEY )

LOAD

START

,

END

,

KEY

RESIDENT

DIMENSION

Solution was : (+ lots of load distinct before )

LEFT JOIN INTERVALMATCH (DATE, KEY) LOAD Distinct START, END, KEY RESIDENT DIMENSION;

LEFT JOIN (FACT) LOAD Distinct START, END, JOIN, Field RESIDENT DIMENSION;

//DROP TABLE KAMP;

DROP FIELDS START, END, KEY FROM FACT;