Hi Scott

The manual shows a sample datetime input format as 11 Nov 2019 13:10:00.  But when I look at the day and month values in your DATE8 and DATE10 columns, I suspect that the input format interprets your 8 digit date as a JULIAN date: 20251202 -> 2025.120 -> April 30, 2025, and it ignores the 8th character.

If you could massage the format of your data set, can you insert dashes between the year, month and day parts?

> Also is there a way to convert a Carla date to this format YYYYMMDD without / or - ?

This request has been made more times than I can count, and never made it to the development list.  Feel free to open an Idea but also check how many Ideas there are already that don't get implemented.  Anyway, you could look at composing the date from 3 parts like

sortlist date10(year) | date10(usdate,2) | date10(eudate,2)

Horribly inefficient because the date is calculated and converted 3 times, but hey, it may get you what you need.  Here is a sample of date output strings. 

------------------------------
Rob van Hoboken
------------------------------

Thanks Rob, yes I did try adding dashes, slashes and spaces in between the Year/Month/Day and each will resolve the dates. 

However I was wondering is there a way I could identify records older than 90 days without converting MyDate (YYYYMMDD) to a date?  I tried a couple of things:

I passed a symbolic jcl value to the carla step and called it THRESH.  I then created these Carla statements but it didn't produce any records, 

SYMBOLIC num thresh=&thresh  <-- In the sysprint this does resolve SYMBOLIC num thresh=20251203 

Select mydate<thresh-90 <- but it didn't identify any records. So this didn't work. 

 So then I tried a select statement referencing the jcl symbolic value:

select mydate<&thresh-90.  And in the sysrpint it shows it did resolve the value "select mydate<20251203-90" but I received CKR0412 12 String longer than expected size 10 at SYSIN line 14.  

Is there a way I could pass today's date as a jcl symbolic value and then use it to do math with in the carla step (e.g. mydate<&thresh-90)?

Hi Scott

The manual shows a sample datetime input format as 11 Nov 2019 13:10:00.  But when I look at the day and month values in your DATE8 and DATE10 columns, I suspect that the input format interprets your 8 digit date as a JULIAN date: 20251202 -> 2025.120 -> April 30, 2025, and it ignores the 8th character.

If you could massage the format of your data set, can you insert dashes between the year, month and day parts?

> Also is there a way to convert a Carla date to this format YYYYMMDD without / or - ?

This request has been made more times than I can count, and never made it to the development list.  Feel free to open an Idea but also check how many Ideas there are already that don't get implemented.  Anyway, you could look at composing the date from 3 parts like

sortlist date10(year) | date10(usdate,2) | date10(eudate,2)

Horribly inefficient because the date is calculated and converted 3 times, but hey, it may get you what you need.  Here is a sample of date output strings. 

------------------------------
Rob van Hoboken
------------------------------

Hi Scott,

CARLa does not have arithmetic operators, with one exception, you can calculate relative dates using TODAY-nn and DUMPDATE-nn.  See here.  The manual states that the left part of the condition must be a date field.

SYMBOLIC was made with one purpose in mind, and that is replacing the text string by some digits.  Not an expression.

You have 2 options:

1. pass the actual threshold value by JCL, so today is December 3rd, 90 days ago was at or around September 5th, so
   // SET CUTOFF=20250905
//SYSIN DD *,SYMBOLS=JCLONLY
...
select mydate<&cutoff
2. you can generate the SELECT statement in a Rexx program that you execute before the C2RC proc, write a NEWLIST block into a data set:
   /* Rexx */
queue 'newlist type=$input outlim=0 name=cutoff'
queue 'select mydate<' || date('S', date('B')-90, 'B')
   Allocate this data set to file CUTOFF of your C2RC, use INCLUDE DD=CUTOFF to load this CARLa block.
   Refer to the result of the newlist with SELECT xxxxxx LIKELIST=CUTOFF

Thanks Rob, yes I did try adding dashes, slashes and spaces in between the Year/Month/Day and each will resolve the dates. 

However I was wondering is there a way I could identify records older than 90 days without converting MyDate (YYYYMMDD) to a date?  I tried a couple of things:

I passed a symbolic jcl value to the carla step and called it THRESH.  I then created these Carla statements but it didn't produce any records, 

SYMBOLIC num thresh=&thresh  <-- In the sysprint this does resolve SYMBOLIC num thresh=20251203 

Select mydate<thresh-90 <- but it didn't identify any records. So this didn't work. 

 So then I tried a select statement referencing the jcl symbolic value:

select mydate<&thresh-90.  And in the sysrpint it shows it did resolve the value "select mydate<20251203-90" but I received CKR0412 12 String longer than expected size 10 at SYSIN line 14.  

Is there a way I could pass today's date as a jcl symbolic value and then use it to do math with in the carla step (e.g. mydate<&thresh-90)?

Hi Scott

The manual shows a sample datetime input format as 11 Nov 2019 13:10:00.  But when I look at the day and month values in your DATE8 and DATE10 columns, I suspect that the input format interprets your 8 digit date as a JULIAN date: 20251202 -> 2025.120 -> April 30, 2025, and it ignores the 8th character.

If you could massage the format of your data set, can you insert dashes between the year, month and day parts?

> Also is there a way to convert a Carla date to this format YYYYMMDD without / or - ?

This request has been made more times than I can count, and never made it to the development list.  Feel free to open an Idea but also check how many Ideas there are already that don't get implemented.  Anyway, you could look at composing the date from 3 parts like

sortlist date10(year) | date10(usdate,2) | date10(eudate,2)

Horribly inefficient because the date is calculated and converted 3 times, but hey, it may get you what you need.  Here is a sample of date output strings. 

------------------------------
Rob van Hoboken
------------------------------