topic Re: How to optimize like this? in QlikView

How to optimize like this?

Tue, 13 Mar 2012 13:21:35 GMT

Hi All,

Actually i did like this.

I wrote the script for getting the data along with the incremented.

But with this i need to reload the data each and every time to get the data along with the incremented.

For optimization i want to do like below.

Now how can i avoid the reload for the alredy loaded data in the qvw file and do the reload for the incremented data if exist.so that i can get the total data.

If i do like this the performance of the application will be increased and reloading time will be decreased.

Is it possible to optimize like this?

Thanks for any help

Regards

.....

How to optimize like this?

jvitantonio — Tue, 13 Mar 2012 13:28:00 GMT

You need to use ADD after the LOAD statement and use the Partial Reload option from the File menu. Check

Ex:

ADD load * from file.qvd (Qvd)

where not exists(field)

Re: How to optimize like this?

Tue, 13 Mar 2012 14:19:52 GMT

Hi,

can you please attach a sample file with this scenario.

Regards

....

Re: How to optimize like this?

sebastiandperei — Tue, 13 Mar 2012 15:27:25 GMT

Hi Rav.

I Think you could think in this example:

Total_Sales:

Load

Date,

Cust,

Sales

From Total_Sales.qvd(qvd); <= This is the qvd that will add the total sales

Let MaxDate = Max(Date);

Concatenate

Load

Date,

Cust,

Sales

From New_Sales.qvd(qvd) <= Here you must consider the data source

Where Date > $(MaxDate); <= Consider in use the fields that have newer values

The performance will depends on the amount of data and their structure. For example, if you can't use Date > $(MaxDate), because of your old data is updated too, you need to use Not Exists, and it take a lot of resources...

Give more information please

Re: How to optimize like this?

Wed, 14 Mar 2012 07:39:45 GMT

Hello Sebastian,

Thanks for your reply.

What i am doing is like that only.

But what i am asking is

ex:

Test:

Load

Date

Group

from test.qvd(qvd);

vMaxdate=Max(Date);

Concatenate

Load

Date,

Group

from test.qvd(qvd) where Date>Max(Date);

If i reload the above script

from the first load statement, i will get the original data

from the second load statement, i will get the increment data if exists otherwise i will get the original data only.

Here i want to store the data exists after the reload in some memory(like assume) and when i reload this script again if there is no changes in the source file, it need's to get back the same data from that memory(assumed) without reloading.So,that i want to avoid the reload for already reloaded same source once.

Now if there is any incremental data.That incremental data only needs to be reload and adds to the data which is in memory(assumed).Now the whole data needs to be store in that memory(assumed).Like wise i am thinking to do.

I don't know is it possible or not.

This is how i want to do.

I think u understand my explanation.

I need help in this if it possible.

Thanks in advance.

Regards

.....

Re: How to optimize like this?

sebastiandperei — Wed, 14 Mar 2012 13:05:42 GMT

In first place, sorry if I don't understand you, my english is not too good.

The place in memory you said is the famous hard disk! After the reload script, you store all the data in Total_Sales.qvd. Then, when you load this file again, at the beginning of the script, you will take again all the last data stored. Pay attention, like this way will function better:

Data:

Load

Date,

Group

From Total_Sales.qvd(qvd);

Aux:

Load Max(Date) as MaxDate

Resident Data;

Let vMaxDate=Peek('MaxDate');

Drop table Aux;

Concatenate (Data)

Load

Date,

Group

From [Source]

When Date > $(MaxDate);

Store Data into Total_Sales.qvd(qvd);

The good notice is that the loading with no modifications (as, when, etc...) is very faster from qvd!.

But in some cases, it doesn't work better than the traditional way.

Re: How to optimize like this?

sebastiandperei — Tue, 20 Mar 2012 16:09:17 GMT

It was useful for you?