TM1 Forum

Hi, I'am new member in this forum, and I am beginner in this field of BI. and sorry for my english null

I want download data from a text file ( log file ) that contains data about the execution of different process and put them in a cube.

In the " log " file, there are a columns of data delimited bye ', ' - for example :

"2","20150417142238","20150417142238","Admin","S","\\Sources\Dimensions\","\\\TM1__LOAD\Sources\Dimensions\","Z_Parametre","DimensionSourceRepository","Svalue",""

so i want test if it's a process , if true i put the time information and duration of process in the cube, so I will just extract data from some column,

so finaly i want have a cube that contains all information about all process execution.

And i want do all this by using TAB PROLOGUE program.

Thank you for help

cordially

Hi

Just I forget to say that i am working in : IBM COGNOS TM1 Architect 10.2

Thank you

Have you designed the cube you will use to hold your data?

That's a nice laundry list of what you want the cube to do. Have you spent any time reading the documentation? Thought about taking a training class? This forum is not really here to walk you step by step through the things that are covered in the documentation. I suggest you spent some time reading before you spend any more time asking.

Thank you for your answer.


David Usherwood : thank you very much for you answer , Yes i have create a cube (named : Z_Suivi_Process) that containt 3 dimensions : Time ( year,trimestre,month) - Cube (containt all process ) - Time_mesure ( start_time, stop_time, duration time , number_of_execution)

so i have write some code line , I start by retrieve the system date and I concatenated with ' TM1S ' to get a string like :TM1S20150513...
and like that I get the log file for today from server(info: the process will always executes in the late afternoon )

And now I 'm stuck How do I get data from specific columns log file and put it in the cube, that's all.


For Tomok : i don't want a solution step by step, i was just asked et now if it's possible to do this just with program Prolog.

I am a beginner and Décisionel IT, i am now in a intership , it's my first intership in this field.

Thankyou

SysHic wrote:...i don't want a solution step by step, i was just asked et now if it's possible to do this just with program Prolog.

No. You can't do this just using the Prolog because the prolog executes in isolation without binding to a data source. To examine the contents of a text file line by line as a data source you need to use the metadata or data tab as code on these tabs will be repeated per record of the data source. On the prolog you would setup your variable declarations, test parameters and set what the data source is going to be.

lotsaram wrote:

SysHic wrote:...i don't want a solution step by step, i was just asked et now if it's possible to do this just with program Prolog.

No. You can't do this just using the Prolog because the prolog executes in isolation without binding to a data source. To examine the contents of a text file line by line as a data source you need to use the metadata or data tab as code on these tabs will be repeated per record of the data source. On the prolog you would setup your variable declarations, test parameters and set what the data source is going to be.

Thank you for your answer. ok i understand . but the probleme here it's a file texte no a CSV text so there are no header. how i call the specific data

i was find this commande : DIMENSIONELEMENTINSERT and DIMENSIONELEMENTCOMPONENTADD , but This function adds an element not sub not all the data. -will you have an idea on the command I should use. thank you so much

It is obvious from your questions you are completely and totally lost. You really need to read the documentation, especially that which covers Turbointegrator. All your questions will be answered (at least most of them anyway). After you've had a chance to digest the docs then come back with specific questions when you run into a brick wall.

Hi SysHic ,Please go through below link and read Alan Kirk and Steve Vincent replies ,It would give understanding of Advance tab.
http://www.tm1forum.com/viewtopic.php?f=3&t=1103

Also,the function you should be looking at is CellPutN or CellPutS to be used in the data tab.As suggested by Tomok,Please go through the TM1 documentation to get understanding of TM1.Thanks

SysHic wrote:Thank you for your answer. ok i understand . but the probleme here it's a file texte no a CSV text so there are no header. how i call the specific data

i was find this commande : DIMENSIONELEMENTINSERT and DIMENSIONELEMENTCOMPONENTADD , but This function adds an element not sub not all the data. -will you have an idea on the command I should use. thank you so much

Once you've setup your data source as a text file in the opening tab you should see the columns within the file (seperated by a comma) in the variables tab. Each variable needs to havea unique name. When you setup your source the TI process tries to name the variable based on the contents of the column. Normally its best to change the name yourself to something meaningful, so for example if your column contains years call it vYear. The prefix of "v" signifies that it is a variables. Then in the data / meta data tabs when you refer to the variable it actually refers to the value of the variable. This is what Lotsaram was meaning. I hope this helps.

BTW Tomok is right. You really do need to at least go through the documentation that IBM provide. There is a developers guide that covers most of what we have pointed out so far. At the moment you're trying to drive a car without getting lessons,

Jim.

jim wood wrote:

SysHic wrote:Thank you for your answer. ok i understand . but the probleme here it's a file texte no a CSV text so there are no header. how i call the specific data

i was find this commande : DIMENSIONELEMENTINSERT and DIMENSIONELEMENTCOMPONENTADD , but This function adds an element not sub not all the data. -will you have an idea on the command I should use. thank you so much

Once you've setup your data source as a text file in the opening tab you should see the columns within the file (seperated by a comma) in the variables tab. Each variable needs to havea unique name. When you setup your source the TI process tries to name the variable based on the contents of the column. Normally its best to change the name yourself to something meaningful, so for example if your column contains years call it vYear. The prefix of "v" signifies that it is a variables. Then in the data / meta data tabs when you refer to the variable it actually refers to the value of the variable. This is what Lotsaram was meaning. I hope this helps.

BTW Tomok is right. You really do need to at least go through the documentation that IBM provide. There is a developers guide that covers most of what we have pointed out so far. At the moment you're trying to drive a car without getting lessons,

Jim.

---------------------------------------------------------------------------------------------------------------------------------------------------------------------
Thank you for the answer. I understand, but the problem that it is a txt file and not a CSV file, there is no header on the columns that make references.

For example, I have a CSV file that contains:

N5 Lib_N5 N4
Total Libelle Total Code
TOT_ETJ Total Entité A120
TOT_ETJ Total Entité A130
TOT_ETJ Total Entité A201
ETJ_TECH Entité - TECH CN
ETJ_TECH Entité - TECH CN
ETJ_TECH Entité - TECH CN
ETJ_TECH Entité - TECH A241

there has a header for 3 columns N5, Lib_N5, N4, so refer to the data, and therefore I can do, for example:

AttrPutS (Lib_N5, VDIM, N5, vDimAlias);

But in my case it is a text file, there is no header, and this is my big probleme. how i can refer the data in my log file.

The fact that the file has no header makes no difference at all which is what I was trying to point out. In fact if it has a header row then you should skip as it's not valid data. The variable names have nothing to do with the headers. The process will pick them up and you specify a header row and the header contains no spaces but you can simply go in to the variables tab and change the variable name. This is covered by the documentation I mentioned earlier. You really need to go through as you clearly don't understand how TM1 refers to variables.

jim wood wrote:The fact that the file has no header makes no difference at all which is what I was trying to point out. In fact if it has a header row then you should skip as it's not valid data. The variable names have nothing to do with the headers. The process will pick them up and you specify a header row and the header contains no spaces but you can simply go in to the variables tab and change the variable name. This is covered by the documentation I mentioned earlier. You really need to go through as you clearly don't understand how TM1 refers to variables.

Ok, thank you , When i can't find this documentation please ? i cannot find it

cordially

Within Perspectives select help then contents and index. You should see a link to the developers guide

jim wood wrote:Within Perspectives select help then contents and index. You should see a link to the developers guide

Thank you very much Sir for all answers

Cordially

You could also find Cognos TM1 version 10.1 and 10.1.1 product documentation below:
http://www-01.ibm.com/support/docview.w ... wg27021893 Thanks