Getting file lines randomly

On Mon, 24 Jul 2006 contracer11@gmail.com wrote: > Is there any way, using a dcl procedure, to select 3 lines in a file , > randomly ? Sure. DCL is a programming language. On the VMS FAQ, you will find a random number generator. You can use SEARCH/STAT (for example) to find out how many lines are in file. You can use the DCL OPEN command to open the file and the READ command to read the file sequentially, discarding lines that you don't want to print. If you had, or were willing to create, unique keys for each record, you could create an indexed file and use READ/KEY and avoid reading the file sequentially. Don't forget to issue the DCL CLOSE command to close the file. -- Rob Brown b r o w n a t g m c l d o t c o m G. Michaels Consulting Ltd. (780)438-9343 (voice) Edmonton (780)437-3367 (FAX) http://gmcl.com/

contracer11@gmail.com wrote: > Is there any way, using a dcl procedure, to select 3 lines in a file , > randomly ? Yes, this selection is easily possible by acquiring any of various values (often derived from the system time and I/O counts and other such factors) and using this to generate an index on a loop that reads in the records, stopping when the pseudo-random index value is reached. Writing some sort of password tool? If so, there are rather better ways to deal with this particular situation. (There are tools and mechanisms that allow rather more random passwords, and there are DCL procedures that use these tools available for examination.) There are various "toys" that use a sequence similar to what you seem to want here, the most common one I can think of has three sets of (usually) ten words, and can be used as a buzz-phrase generator. For example, and realize I have no connection with any of the following two URLs: http://www.acronymfinder.com/buzzgen.asp http://www.cafepress.com/iqthereforeiam.35260768 Buzz-phrase bingo is a pursuit encountered occasionally in various presentations, as well. Whether subversive or a survival instinct depends on whether you're calling the game or playing it, of course.

Rob Brown wrote: > On Mon, 24 Jul 2006 contracer11@gmail.com wrote: > > > Is there any way, using a dcl procedure, to select 3 lines in a file , > > randomly ? > > Sure. DCL is a programming language. > > On the VMS FAQ, you will find a random number generator. > > You can use SEARCH/STAT (for example) to find out how many lines are > in file. > > You can use the DCL OPEN command to open the file and the READ command > to read the file sequentially, discarding lines that you don't want to > print. > > If you had, or were willing to create, unique keys for each record, > you could create an indexed file and use READ/KEY and avoid reading > the file sequentially. > > Don't forget to issue the DCL CLOSE command to close the file. > > > -- > > Rob Brown b r o w n a t g m c l d o t c o m > G. Michaels Consulting Ltd. (780)438-9343 (voice) > Edmonton (780)437-3367 (FAX) > http://gmcl.com/ What I am doing wrong ?? procedure below only show me file first line... Alpha>> ty ap.com $ set nover $ now_time = F$time() $ hrs = F$cvtime(now_time,, "hour") $ min = F$cvtime(now_time,, "minute") $ sec = F$cvtime(now_time,, "second") $ hun = F$cvtime(now_time,, "hundredth") $ $ ! $ ! Get a good seed to start the madness. $ ! $ seed = 360000 * hrs + 6000 * min + 100 * sec + hun $ seed = seed .and. %XFFFF $ $ highval = 100 ! Get a number from 0 to 100. $ gosub _RANDOM ! Get the random $! write sys$output random $ open/read file users.lis $ read/index='random' file registro $ write sys$output "''registro'" $ close file $exit $_RANDOM: $ seed = (seed * 69069 + 1) .and. %XFFFFFFFF $ rand = (seed / %X100) .and. %XFFFF $ t = rand / highval $ k = %XFFFF / highval $ if rand .gt. (k * highval) then goto _RANDOM $ random = rand - t * highval + 1 $ $ Return $ open/read file users.lis $ read/index='random' file registro $ write sys$output "''registro'" $ close file Alpha>> @ap OpenVMS User Processes at 24-JUL-2006 11:24:22.57 Alpha>> @ap OpenVMS User Processes at 24-JUL-2006 11:24:22.57 Alpha>> @ap OpenVMS User Processes at 24-JUL-2006 11:24:22.57 Alpha>>

On Mon, 24 Jul 2006 contracer11@gmail.com wrote: > What I am doing wrong ?? > procedure below only show me file first line... > > Alpha>> ty ap.com > ... > $ open/read file users.lis > $ read/index='random' file registro > $ write sys$output "''registro'" If you created users.lis by SHOW USERS/OUTPUT=USERS.LIS, then users.lis is *not* an indexed file and you are limited to reading the file sequentially. In other words, /INDEX and /KEY will not work. Since it appears that you want to randomly pick some logged on users by line number, it would not seem efficient to create an indexed file with line numbers. Other things are possible, but I think you might as well just read the file sequentially, counting records, and selecting the records that match your random numbers. -- Rob Brown b r o w n a t g m c l d o t c o m G. Michaels Consulting Ltd. (780)438-9343 (voice) Edmonton (780)437-3367 (FAX) http://gmcl.com/

contracer11@gmail.com wrote: > Rob Brown wrote: > > On Mon, 24 Jul 2006 contracer11@gmail.com wrote: > > > > > Is there any way, using a dcl procedure, to select 3 lines in a file , > > > randomly ? > > > > Sure. DCL is a programming language. > > > > On the VMS FAQ, you will find a random number generator. > > > > You can use SEARCH/STAT (for example) to find out how many lines are > > in file. > > > > You can use the DCL OPEN command to open the file and the READ command > > to read the file sequentially, discarding lines that you don't want to > > print. > > > > If you had, or were willing to create, unique keys for each record, > > you could create an indexed file and use READ/KEY and avoid reading > > the file sequentially. > > > > Don't forget to issue the DCL CLOSE command to close the file. > > > > What I am doing wrong ?? > procedure below only show me file first line... > > Alpha>> ty ap.com > $ set nover > $ now_time = F$time() > $ hrs = F$cvtime(now_time,, "hour") > $ min = F$cvtime(now_time,, "minute") > $ sec = F$cvtime(now_time,, "second") > $ hun = F$cvtime(now_time,, "hundredth") > $ > $ ! > $ ! Get a good seed to start the madness. > $ ! > $ seed = 360000 * hrs + 6000 * min + 100 * sec + hun > $ seed = seed .and. %XFFFF > $ > $ highval = 100 ! Get a number from 0 to 100. Are there always 100 records in the file? If not, you need the exact count and since you need to read through the file anyway: $ highval = 0 $ open/read file users.lis $! $loop1: $ read/end=counted file registro $ highval = highval + 1 $ save'highval' = registro $ goto loop1 $! $counted: $ close file $ if highval .gt. 0 $ then $ count = 0 $ loop2: $ gosub _RANDOM ! Get the random $ if random .gt. 0 - .and. random .le. highval ! paranoia test $ then $ write sys$output save'random' $ count = count + 1 $ endif $ if count .lt. 3 then goto loop2 $ else $ write sys$output "File is empty." $ endif $exit $! $_RANDOM: $... Haven't really tested your _RANDOM routine, thus the paranoia test. Others have explained the read/index problem.

contracer11@gmail.com wrote: > > $ read/index='random' file registro To repeat, /INDEX is used for indexed (ISAM) files, only. It is meaningless for other file formats. See http://www.djesys.com/vms/mentor/rms.html for information about RMS file formats. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/