Deduplication of Rows — Oracle Analytics

Oracle Analytics Cloud and Server

Welcome to the Oracle Analytics Community: Please complete your User Profile and upload your Profile Picture

Deduplication of Rows

Received Response
1
Views
3
Comments
user5388961
user5388961 Rank 1 - Community Starter

Hi,

I have the following columns and the situation.

Total Customers: (Count distinct of cust no)

                                  channel    Values below

                                  Internet =  1 , null

                                  branch =   1, null

                                  mobile =   1, null

my report is

1) month,  total customers

201510 7000000

when i run another report with Channel Type (case statement) with Month and Total Customers

case statement for channel type column

case when internet =1 then internet

when branch =1 then branch

mobile =1 then mobile

else others end

2) month     channel type     total customers

201510    internet               3600000

201510     branch               2800000

201510    mobile                  600000

3) but if i choose total customers where branch =1 then it is giving 4400000, as the 2nd report is giving unique customers of  each channel

where as the 3 rd report gives the customers who use branch channel, these customers might also use Internet and mobile

if i want to get 4400000 for branch channels, how can i do this. if i choose the channel type attribute the column count the customers only once in any of the channel. but i want to count each customer once if  customer use 3 channels.

please let me know any ideas.

Answers

  • Azhar01
    Azhar01 Rank 1 - Community Starter

    Hi     Is the case statement correct. ?

    case when internet =1 then internet

    when branch =1 then branch

    mobile =1 then mobile

    else others end

    or it should be

    case when channel _type =1 then internet .... like that

  • Thomas Dodds
    Thomas Dodds Rank 8 - Analytics Strategist

    Perhaps a dedicated fact table that has that count already figured out - it's much more efficient.

  • [Deleted User]
    [Deleted User] Rank 2 - Community Beginner

    +1 to Thomas. It's much more efficient and you have a lot more possibilities for derived calculations which will otherwise yield "wrong" results with regards to pre/post-aggregate calculations.

    Also all dimensionalities around it will automatically work and aggregate.