Last year I wrote about the similarity and superiority of the SQR **evaluate** command compared to the C or Java **switch** command in SQR Evaluate Versus C/Java Switch, part 1 and part 2. Here is another way to extend **evaluate** functionality.

### Evaluate Two Values

**Evaluate** takes one value and compares it to one or more other values. Sometimes we need to perform different operations depending on two values. Consider two columns of the PS_EMPLOYEES table, FULL_PART_TIME, which can be “F” or “P”, and REG_TEMP, which can be “R” or “T.” There are four combinations of values that might require four different actions. We can control program flow like this:

`if &FULL_PART_TIME = 'F'`

if ®_TEMP = 'R'

! branch 1

else

! branch 2

end-if

else

if ®_TEMP = 'R'

! branch 3

else

!branch 4

end-if

end-if

or not improve it like this:

`evaluate &FULL_PART_TIME`

when = 'F'

evaluate ®_TEMP

when = 'R'

! branch 1

when = 'T'

! branch 2

end-evaluate

when = 'P'

evaluate ®_TEMP

when = 'R'

! branch 3

when = 'T'

! branch 4

end-evaluate

end-evaluate

We can remove a level of nesting and complexity with this:

`let $key = &FULL_PART_TIME || ®_TEMP`

evaluate $key

when = 'FR'

! branch 1

when = 'FT'

! branch 2

when = 'PR'

! branch 3

when = 'PT'

! branch 4

when-other

! invalid values

end-evaluate

### When-Other, Whatever

I’m confident that these two columns from the PS_JOB table, via the PS_EMPLOYEES table, will have values (Peoplesoft character fields are always defined as non-null) and they will have the predicted values (as enforced by the PIA page processor). I use a **when-other** branch because I don’t want the program to proceed heedlessly when it turns out that I was overconfident. How could that happen?

- Someone updates a PS_JOB or PS_EMPLOYEES row in SQL, bypassing Peoplesoft validation.
- The
**where**clause of my PS_EMPLOYEES selection doesn’t return any rows, leaving the column variables set to null. - The latest release of Peoplesoft HCM introduces new valid values for these columns.

### About This Post

Last time I wrote about **evaluate**, it took 2400 words and two weeks to express all my thoughts. A while later, I had the afterthought you’ve just read. I saved it, waiting until I had enough material for a full sized post. I did have more to write about **evaluate**, but it was in the nature of “here’s a farfetched use of **evaluate**, but it would be unwise to do it.” As a result, I withheld this idea too. This year, I will allow myself to publish shorter posts, which will give us new topics on which I have a bit, but not a lot, to say.

## Recent Comments