Join Data from Excel

[#1]

Hello everybody

I have a problem with a load script

I have a fact table with all the information and I need to add a colunm from a Excel File. The problem is that some of the data in the table not apear in the excel File. Example

Fact table (prior Join)

ITEM

00001

00002

00003    

00004

00005

Excel

ITEM          ITEM_KEY

00001          101001   

00004          101004

00005          101005

Fact table (With a simple Join)

ITEM          ITEM_KEY

00001          101001

00002          -

00003          -   

00004          101004

00005          101005

What I need is, if the value item is in the exel (relation table) the field in ITEM_KEY needs to be equal to the value in the excel, but if the value is not in the Excel the Load need to pass the value of the ITEM column. The result I need for the load is the next:

ITEM           ITEM_KEY

00001          101001 (This value is from the Excel)

00002           00002 (This value is from the Item column form the same fact table)

00003           00003 (This value is from the Item column form the same fact table)

00004          101004 (This value is from the Excel)

00005          101005 (This value is from the Excel)

Can anyone could help me

[joshabbott]

In the excel spreadsheet load, do a preceding load on top of it.  For example, your load may look something like:

Load

  ITEM

  ,ITEM_KEY

from excel;

Try this:

Load

  ITEM

  ,Alt(ITEM_KEY, ITEM) as ITEM_KEY

Load

  ITEM

  ,ITEM_KEY

from excel;

The alt should get the first non-null value and return it to the field.

[joshabbott]

Better yet, you may not need to do the preceding load at all, instead of LOAD ITEM_KEY

Try:

LOAD

  Alt(ITEM_KEY,ITEM) as ITEM_KEY

[maxgro]

you can use a mapping load

Excel:

mapping load * inline [

ITEM      ,    ITEM_KEY

00001     ,     101001  

00004     ,     101004

00005     ,     101005

];

[Fact table]:

load

  ITEM,

  ApplyMap('Excel', ITEM, 'Value not found ' & ITEM) as ITEM_KEY;

load * inline [

ITEM

00001

00002

00003   

00004

00005

];