Help with Load Script

[#1]

I have a list of customers which I need to segregate based on the unique customers in each channels which comes from an excel file, refer the above screenshot. I have this built in chart but as this requires aggregation of millions of rows chart ends up eating the entire RAM and make the app very slow, please help.

I want customer 2 and 3 to be marked as Phone_Only and 5 as Web only and remaining as Web and Phone.

Thanks for your help!

Nidhish

[sunny_talwar · Accepted solution]

May be something like this:

Script:

Table:

LOAD * Inline [

Cus, Channel

1, Phone

1, Web

2, Phone

3, Phone

4, Phone

4, Web

5, Web

];

Join(Table)

LOAD Distinct Cus,

  Concat(Channel, '|') as GroupedChannel

Resident Table

Group By Cus;

FinalTable:

LOAD Cus as Customer_Phone_Only,

  Channel

Resident Table

Where WildMatch(GroupedChannel, '*Phone*') and not WildMatch(GroupedChannel, '*Web*');

Concatenate (FinalTable)

LOAD Cus as Customer_Web_Only,

  Channel

Resident Table

Where not WildMatch(GroupedChannel, '*Phone*') and WildMatch(GroupedChannel, '*Web*');

Concatenate (FinalTable)

LOAD Cus as Customer_Web_and_Phone,

  Channel

Resident Table

Where WildMatch(GroupedChannel, '*Phone*') and WildMatch(GroupedChannel, '*Web*');

DROP Table Table;

[crystles]

Can you provide a picture of the table structure from the table view? How are the three tables linked?

[Anonymous]

You may need to sort this in your load script to get a flag in your fact table for :

• Phone Only
• Web Only
• Both

Then you won't need any front end aggregations.

[#5]

This worked, and I was able to find one more workaround, thanks a lot for the quick help!