TM1 Forum

Hi All ,

Our Budget cubes are having many rules , feeders and calculations . So opening the view is taking more than 10 minutes . To overcome this we have created dummy cubes which has no rules and feeders . All the data in the dummy cubes are zeroed out and is loaded from Budget cubes via scheduled TI process .This scheduled process is running for 15 minutes & bussiness needs to review the changes in one hour .

Running the scheduled chore for every hour is affecting users who logs in for Budgeting the numbers. I want to know whether it is possible to access the transaction log of tm1 server through a TI process so that only the changes which occured can be zeroed out and loaded into the dummy cubes ?

Any suggestions or ideas are most welcome .


Thanks
Anoop

The transaction log is just a text file so it can certainly be opened and used in a TI process. However, you can't open the "current" log file because it is going to be locked by the TM1 service and thus not available for opening. You'll have to do a SaveDataAll first to commit those changes to disk, at which time the "current" log file is closed and a new one started. The challenge here is going to be specifying exactly which file(s) to open automatically. When the log file is closed it is renamed with a GMT date and time stamp. Perhaps you can write some sort of batch file in WSH that will rename the "unprocessesed" log files, or better yet copy them, to a name your TI process can rely on. Good luck.

Savedataall() ; Allows you to close off the current transaction log and write the contents to an archive transaction log with the date time stamp on it. This is no longer locked by TM1 and is accessible through programming as a .CSV file.

You would need some sort of dir *.log > temp.txt and search temp.txt for the latest log and then rename it to something standard so a TI proc could read it.

Savedataall() ; Allows you to close off the current transaction log and write the contents to an archive transaction log with the date time stamp on it. This is no longer locked by TM1 and is accessible through programming as a .CSV file.

You would need some sort of dir *.log > temp.txt and search temp.txt for the latest log and then rename it to something standard so a TI proc could read it.

Whoops - great minds think alike tomok - post just after yours!

anoops81 wrote:Hi All ,

Our Budget cubes are having many rules , feeders and calculations . So opening the view is taking more than 10 minutes . To overcome this we have created dummy cubes which has no rules and feeders . All the data in the dummy cubes are zeroed out and is loaded from Budget cubes via scheduled TI process .This scheduled process is running for 15 minutes & bussiness needs to review the changes in one hour .

Running the scheduled chore for every hour is affecting users who logs in for Budgeting the numbers. I want to know whether it is possible to access the transaction log of tm1 server through a TI process so that only the changes which occured can be zeroed out and loaded into the dummy cubes ?

Any suggestions or ideas are most welcome .


Thanks
Anoop

Yes it's possible to process transaction logs (and I suspect most consultants would have generic TI processes to do this). However you are still going to need to know WHICH log to process or have a mechanism in place for processing all logs for a specified time window. As already noted above you can't process the current transaction log as the TM1 server has a lock on the file. (Doing a save data to get around this could also have its own performance implications if a lot of data has changed.)

However I think this misses the point. Is the intention merely to review WHO has made changes and WHERE or to do a QUALITATIVE and QUANTITATIVE assessment of the significance of the changes? I would assume or rather hope that it is the later and not the former as a mere count of the number of changes adds no value other than a review for its own sake.

If this is the case then the transaction log is not going to help you! you have stated that the budget cubes use a lot of rules (which presumably calculate the impact in financial terms of driver inputs and assumptions), so the answer to the important question the CFO wants answered "what has changed?" lies in the end result of those calculations and not in the data input that you can pull out of the transaction log.

lotsaram wrote:

anoops81 wrote:Hi All ,

Our Budget cubes are having many rules , feeders and calculations . So opening the view is taking more than 10 minutes . To overcome this we have created dummy cubes which has no rules and feeders . All the data in the dummy cubes are zeroed out and is loaded from Budget cubes via scheduled TI process .This scheduled process is running for 15 minutes & bussiness needs to review the changes in one hour .

Running the scheduled chore for every hour is affecting users who logs in for Budgeting the numbers. I want to know whether it is possible to access the transaction log of tm1 server through a TI process so that only the changes which occured can be zeroed out and loaded into the dummy cubes ?

Any suggestions or ideas are most welcome .


Thanks
Anoop

Yes it's possible to process transaction logs (and I suspect most consultants would have generic TI processes to do this). However you are still going to need to know WHICH log to process or have a mechanism in place for processing all logs for a specified time window. As already noted above you can't process the current transaction log as the TM1 server has a lock on the file. (Doing a save data to get around this could also have its own performance implications if a lot of data has changed.)

However I think this misses the point. Is the intention merely to review WHO has made changes and WHERE or to do a QUALITATIVE and QUANTITATIVE assessment of the significance of the changes? I would assume or rather hope that it is the later and not the former as a mere count of the number of changes adds no value other than a review for its own sake.

If this is the case then the transaction log is not going to help you! you have stated that the budget cubes use a lot of rules (which presumably calculate the impact in financial terms of driver inputs and assumptions), so the answer to the important question the CFO wants answered "what has changed?" lies in the end result of those calculations and not in the data input that you can pull out of the transaction log.

Hi Lotsaram ,

The only intention for accessing the transaction log is to load all the transactions (if possible) which occured in the budget cubes for a particular time period to the dummy cube so that the time required for loading the complete data (zeroing out dummy cube completely and loading again from Budget cubes) in the dummy cube can be minimized by loading the transactions alone to the dummy cube .

Thank you all for the ideas which you have provided . I will have to plan for a "brain storm" session on how to get the transaction data in a standard format from the transaction log so that TI process can access it.

Thanks
Anoop