Interval Match taking up too much memory

[sifatnabil]

Hi all,

I used Henric's Interval match method without joins, but my event table has 1 million rows and my date table has 400,000 rows. As a result, during load 70 million intervals were generated. After the load is done, the application hangs (not enough memory). This is my data model below:

Is there a way round this? I don't actually need the intervals to stay in the load. My event table has a numeric CUSTNO, and I need to map the NAME linked to that CUSTNO. However, a single CUSTNO is linked to different NAMEs at different periods.

My dates table:

CUSTNO	NAME	FROM_DATE	TO_DATE
15431	Bank A	20150144	99999999
87438	Bank B	20140101	20150104

My fact table (and the NAME values that I need mapped):

CUSTNO	EVENT	MSG_DATE_NUM	NAME VALUE THAT I NEED
15431	AAAA	20140801	Bank C
15431	BBBB	20150201	Bank A

Is there a way to get rid of the whole intervals and just get the NAME values I need? Otherwise, my QlikView is hanging. Thanks.

[danieloberbilli · Accepted solution]

Have a look at this document IntervalMatch and Slowly Changing Dimensions : https://community.qlik.com/docs/DOC-4310

There are concepts of how to get rid of the bridgetable and how to join to the fact table

[sifatnabil]

Thanks for the resources. The issue was on my end where I had far too many open intervals that were completely unnecessary.