Re: dfsort - microseconds (STCK/TOD)

Hi Kolusu, Now everything looks perfect and I share the relevant part of your solution happily: ... C'"TRNSTCK":"', 393,8,DC1,EDIT=(-TT-TT), C'T', 393,8,TC1,

Re: dfsort - microseconds (STCK/TOD)

On Mon, 19 Sep 2022 19:23:36 +, Sri h Kolusu wrote: > >STCKE is only guaranteed to have UNIQUE within SYSPLEX. So, if the >transactions are across multiple sysplexs then you are not guaranteed to have >unique values. > For uniqueness µsec granularity is insufficient; TOD can return 4100 dif

Re: dfsort - microseconds (STCK/TOD)

>> Ooooh! ISO 8601! Is that well supported by DFSORT? I don't see it >> mentioned in the Guide. Paul, DFSORT is NOT generating the time stamp in ISO8601 format. It is how OP is formatting the STCKE value with hard coded "T" for time and the UTC offset. Thanks, Kolusu DFSORT Development IBM

Re: dfsort - microseconds (STCK/TOD)

>> Sorry, I thought it to be clear. This are IMS transactions timestamps, >> occurring every x seconds or so: Stefan, STCKE is only guaranteed to have UNIQUE within SYSPLEX. So, if the transactions are across multiple sysplexs then you are not guaranteed to have unique values. How about comp

Re: dfsort - microseconds (STCK/TOD)

On Mon, 19 Sep 2022 01:43:56 -0500, Stefan Lezzi wrote: >... >2022-09-12T13:55:27.836152+:00 >2022-09-12T13:55:27.836152+:00 >... Ooooh! ISO 8601! Is that well supported by DFSORT? I don't see it mentioned in the Guide. Of course for keys they're simply strings.

Re: dfsort - microseconds (STCK/TOD)

Kolusu, Sorry, I thought it to be clear. This are IMS transactions timestamps, occurring every x seconds or so: ... 2022-09-12T13:55:27.836152+:00 2022-09-12T13:55:27.836152+:00 ... 2022-09-12T13:57:28.836152+:00 2022-09-12T13:57:29.836152+:00 2022-09-12T13:57:30.836152+:00 20

Re: dfsort - microseconds (STCK/TOD)

>> but seems still not to be correct: Stefan, You need to explain it a bit more when you say the results are NOT correct. Is it the duplicated data ? or the value itself is wrong? How are you verifying the results? Dump the contents of position 393,8 in hex for may be 20-30 records and then we

Re: dfsort - microseconds (STCK/TOD)

Hi Kolusu, Now I get this, which looks better, but seams still not to be correct: ... 2022-09-12T13:55:27.836152+:00 2022-09-12T13:55:27.836152+:00 ... 2022-09-12T13:57:28.836152+:00 2022-09-12T13:57:29.836152+:00 2022-09-12T13:57:30.836152+:00 2022-09-12T13:57:59.836153+:0

Re: dfsort - microseconds (STCK/TOD)

Stefan, I just realized that you are actually missing the arithmetic operator MOD (which gives the remainder after the divide). I also realized that you have TEMP fields(unhex of TOD) in between which you really don’t need in the final output. So I put the temp fields at the end of build (posi

Re: dfsort - microseconds (STCK/TOD)

data Thanks, Kolusu DFSORT Development IBM Corporation -Original Message- From: IBM Mainframe Discussion List On Behalf Of Stefan Lezzi Sent: Saturday, September 17, 2022 7:32 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] Re: dfsort - microseconds (STCK/TOD) That's a kind offer

Re: dfsort - microseconds (STCK/TOD)

That's a kind offer, but please don't look at the (un)elegance of the code. This made a different thread :-) Just search for TRNSTCK and OVERLAY, at the end. //TOOLINDD * COPY FROM(IN)USING(CTL1) COPY FROM(OUT) TO(OUT2) U

Re: dfsort - microseconds (STCK/TOD)

>> But to be honest, I don't get the math right (positions/offsets/length). I >> always end up with something like this: Stefan, Looking at the output data you have shown, I am guessing that you used OVERLAY to write the time stamp and you have the existing data still in their original positio

Re: dfsort - microseconds (STCK/TOD)

I was prepared to be challenged by such questions :-) Anyway, for my purpose it's not yet so important (and a I can always "correct" the time in a splunk search). Anyway II, I suppose that, here again, I needed deep dfsort (lab) knowledge, if possible at all, to have CVTLSO and CVTLDTO available

Re: dfsort - microseconds (STCK/TOD)

On Fri, 16 Sep 2022 11:18:14 -0500, Stefan Lezzi wrote: > >I want to format a STCK/TOD with microseconds, but I found only the method to >format it with hundredths of a second: > >C'"TRNSTCK":"',393,8,DC1,EDIT=(-TT-TT),C'T', > 393,8,TC4,EDIT=(TT:TT:TT.TT),C'+0

Re: dfsort - microseconds (STCK/TOD)

Hi Kolusu, That's perfectly what I was looking for, many thanks! But to be honest, I don't get the math right (positions/offsets/length). I always end up with something like this: --+5+6+7+8+9+0+ 2022-09-12T13:55:27. 3871979727001140 008B88

Re: dfsort - microseconds (STCK/TOD)

>> I want to format a STCK/TOD with microseconds, but I found only the method >> to format it with hundredths of a second: Stefan, Here is a brief description of how to get the microseconds portion from TOD using DFSORT. Since you haven’t provided the complete DFSORT cards, I assumed that you

Re: dfsort - microseconds (STCK/TOD)

EDU] On Behalf Of Stefan Lezzi Sent: Friday, September 16, 2022 12:33 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: dfsort - microseconds (STCK/TOD) Thanks, I've just tested your suggestion: TT:TT:TT.TT -> 14:53:21.36 TT:TT:TT.TT -> 00:00:14.532136 It behaves like in the man

Re: dfsort - microseconds (STCK/TOD)

Thanks, I've just tested your suggestion: TT:TT:TT.TT -> 14:53:21.36 TT:TT:TT.TT -> 00:00:14.532136 It behaves like in the manual described: Format Code | Length | Description TC4 |8 bytes | TOD time interpreted as Z'hhmmssxx' and "The implied length of the edited o

Re: dfsort - microseconds (STCK/TOD)

What happens if you change TT:TT:TT.TT to TT:TT:TT.TTT? Charles -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Stefan Lezzi Sent: Friday, September 16, 2022 10:15 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: dfsort - microseconds

Re: dfsort - microseconds (STCK/TOD)

a custom HLASM program to produce the desired results. Charles -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Tom Harper Sent: Friday, September 16, 2022 9:41 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: dfsort - microseconds (STCK/

Re: dfsort - microseconds (STCK/TOD)

Hi Tom, Yes, I know, if I would parse the IMS log records with assembler. But assembler is not my strongest skill, and, sadly, not very much loved by company management nowadays. It's about my, already coded and working, dfsort solution to parse some fields of the Mainview/IMS record x'FA' and

Re: dfsort - microseconds (STCK/TOD)

Better to use: https://www.ibm.com/docs/en/zos/2.1.0?topic=routine-description Tom Harper Phoenix Software International Sent from my iPhone > On Sep 16, 2022, at 12:28 PM, Stefan Lezzi > <04561630d35b-dmarc-requ...@listserv.ua.edu> wrote: > > Hi all! > > What do I miss? > > I want t

dfsort - microseconds (STCK/TOD)

Hi all! What do I miss? I want to format a STCK/TOD with microseconds, but I found only the method to format it with hundredths of a second: C'"TRNSTCK":"',393,8,DC1,EDIT=(-TT-TT),C'T', 393,8,TC4,EDIT=(TT:TT:TT.TT),C'+00:00',C'",', And I know how to calcu