Do a "Find" across multiple versions at one time

[DDP021]

I had posted this question in another topic but thought maybe would be better to start a new topic...We have a user loading multiple .txt documents to CMOD..These aren't mainframe text files (with page breaks) but true .txt files (ex test.txt)...We are using a generic indexer to load the documents..The only field the user has on their index file (.ind) is for Posting Date...We have the versions in CMOD open up in WORD (view information on application set to DOC)..What the user would like to do is be able to do "FIND" across multiple versions...The way the Find works now is it only looks at the current version that is opened..I tried select multiple versions in Webi but it only allows you to view on version at a time so when doing the "FIND" it only searches in that version...Does anyone know how to do a FIND across multiple versions at one time using this type of data?...The way they do their searches we don't think indexing will work...

[demaya]

How do you define VERSIONS? I don't understand your definition.

Do you mean applications (like DOC, TXT,...) and you want to search them at one time?

cheers

[DDP021]

Please see the attachment....When I click on the check box next  to the Word icon, it opens up seperate tabs for each check box...It then prompts to open, save or save as the first "version" that was checked..When you close it, you get prompted to do the same with the next "version" that was checked...The issue is, the user would like to open up multiple "versions" at once and then be able to do a find across ALL the versions to bring up all the results...

[Justin Derrick]

If the report type wasnt' word, you might be able to use server-side search at the Folder level to accomplish your goals.  I think the only way to do this is full text searching on CMOD v9.

[DDP021]

Thanks as always Derek!...We tested initially with the user and they at first liked the data opening up in WORD so we set the VIEW INFORMATION on the Application to be DOC for File Extension..I first tried highlighting multiple versions and then selecting VIEW ALL SELECTED (via thick client)...It opened the 3 version in separate  WORD windows...The find though was restricted to only the version that I was in and did the FIND in.....I just tried changing the Application setup from DOC to TXT...I tried the same test and it once again..This time each version in a separate NOTEPAD window...Once again the FIND was restricted to only the window that the find was done in...Was looking for a way, if they selected multiple versions they could then do a FIND and it look across them all..I attached a screen shot showing how the highlighted versions are opening in two separate windows after clicking on the VIEW SELECTED tab...The clients will be using WEBI which, when selecting multiple versions, opens each version in tab

[Justin Derrick]

Yup, I get the requirement -- unfortunately, you'd need to open the document in a program that supported such a feature.  I have a text editor that allows a 'Search all Open Windows', but it's on the Mac. 

-JD.

[DDP021]

Ya, I tried TXT instead of DOC on the Application which resulted in all the windows opening in NOTEPAD...This of course though resulted in the same results...Not sure of any editor to use...These users are located in Germany...As it stands now, they are try to determine if they want to go further with using CMOD to archive their data since the current search doesn't really fill their needs..I was trying to find something that could be used or be done to help them get the results they are looking for...As always, appreciate all you input..Take care....

[Justin Derrick]

If you can load this data as a line data report, you can definitely use server-side search, which is included in the product.  I'm not sure where the Word Docs are coming from, but loading them as text instead of a user-defined format definitely solves your problem.

-JD.

[Greg Ira]

I can't vouch for how it would work through WEBI but what I've done in the past is basically what JD said to try.  I've defined the view for text files as Line data, Code Page 850, Stream ,  CC No.  When these are opened they open with the OnDemand viewer and not notepad(or whatever you have .txt set to on your PC).  With "View Combined documents" option turned on I can select multiple reports and view them as a single document allowing me to search through all the documents using the client based search as opposed to a server side search.

[DDP021]

Thanks Greg for the info as well....These users are located in Germany so the data is also in German...Their only option is to use the WEBI interface...I know you can open multiple versions in webi which then opens separate "tabs" in webi for each version..It however only allows you to open one "tab" at a time to view them...The issue is we have them opening in WORD which of course opens multiple word documents and doesn't allow them to do a "find" across all of them at one time..Only the current Word document they are viewing...These file are located on their TOPAZ (?) system..Not sure what TOPAZ is...They LOOK like they have carriage control (1 in column1 for top of page)..But we didn't know of any way for them to get them from their server to the CMOD server other than sending them as a .txt file and defining the CMOD definition with the GENERIC Indexer and setting the default view as DOC...After loading a bunch of files we find out they want to do the 'searches' across all the versions...If we could 'index' the data like we do with normal line data, using ACIF indexer, that would be best...But we're not sure how the data needs to  be sent/formatted from the user from their server to ours so it recognizes carriage control and we can defined the CMOD definitions using the ACIF indexer as opposed to using the GENERIC INDEXER...

[Greg Ira]

I've indexed certain *.txt files using ACIF with the following parms.  It didn't have a CC in column 1 so we had to find something else but if yours does it would be an easy change. It was largely trial and error getting it to work but I'm thinking you should be able to get it working.
CC=NO
CCTYPE=Z
CONVERT=NO
CPGID=850
MCF2REF=CPCS
LINECNT=80
TRC=NO
FILEFORMAT=STREAM,(NEWLINE=X'0D0A')
TRIGGER1=*,1,X'4F6E',(TYPE=GROUP)                                   /* On           */
FIELD1=0,1,15,(TRIGGER=1,BASE=0)
FIELD2=0,18,10,(TRIGGER=1,BASE=0)
INDEX1=X'5265706F72745F5469746C65',FIELD1,(TYPE=GROUP,BREAK=NO)     /* Report_Title */
INDEX2=X'5265706F72745F44617465',FIELD2,(TYPE=GROUP,BREAK=YES)      /* Report_Date  */
DCFPAGENAMES=NO
UNIQUEBNGS=YES
GROUPMAXPAGES=1
IMAGEOUT=ASIS
INDEXOBJ=ALL
INDEXSTARTBY=1
INSERTIMM=NO
RESTYPE=NONE

[DDP021]

Thanks Greg...This is interesting....We were under the assumption that only line data could use the ACIF indexer.   .txt files would have to use the Generic Indexer...I've attached a sample of one type of file they are trying to load to CMOD..Could the example ACIF indexer you attached in your response be applied to this type of file I've attached?...I don't mean EXACTLY because the Trigger and Index values would obviously be different...Question on a few of your parameters:

FILEFORMAT=STREAM,(NEWLINE=X'0D0A')  < would this value be the same for any .txt file?

TRIGGER1=*,1,X'4F6E',(TYPE=GROUP)  < How did you determine this value?...Comments indicate On but the hex value doesn't...

FIELD1=0,1,15,(TRIGGER=1,BASE=0)  < Curious with a .txt file how you determine the values for the location?..I understand 0 indicates the same line as the trigger...Normally the second parameter (18) would indicate the column..But there are no "column" numbers (like a mainframe line data file) so is the 18 indicating how many spaces AFTER the end of the trigger?...The 10 I know would indicate the length..

For the LINECNT parameter, you have 80...This would indicate the total number of lines per page, correct?..This would have to be consistant, correct?..with a line data file having carriage control, obviously the line count can very...

I tried running the file through the "Report Wizard" to obtain my Trigger and Index values...What I wasn't sure was what to set the NEW LINE VALUE to...I used 0D0C...Set the line count at 80...At this point just trying to get it to load...


I attempted to load the file using ARSLOAD....But it failed with the following error:

Unable to read from offset 0 for length 9320 from the file >/prod/ode/cmod/arsload3/A.B.DAVE3.DAVE3.C.D.ARD<

sorry for all the questions...This is something totally unique from what we've ever done...Appreciate any help or input...

I've attached once sample of one of the files...

Take care

[Alessandro Perucchi]

Hello DDP021,

you must consider Line-data as "text", and by text I mean everything that is considered a text file with its own codepage, (ASCII, EBCDIC, UTF-8, WIN1252, ISO8859-1, ISO8..., ... :-D)
As long as you configure the Application correctly with all the specifics of the text you are reading, then you can use the ACIF Indexer as you want!

Sincerely yours,
Alessandro