![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
My process includes following processes for various clients. - loading excel file into a sql database - Processing loaded data in sql tables - Dumping excel reports I implemented these three stages by using set of physical temporary sql tables (not the memory ones with # and ##) for data transformations etc. I used to delete the data in these temporary tables and reuse for the next client. Till now I was the only person taking care of complete thing. Now I want to add multi-user capability to the system. I feel the only way to implement this is to change all my physical temporary sql tables to memory temporary tables i.e. using (# ## tables). But sql statement in data transformation task does not takes # tables. So I will have to dynamically change the sql to have # tables. Is it ok to load 20k records into memory for each client at the same time. Or is it a good idea to use physical temporary tables i.e. seperate physical table for each client and delteting it at the end. Also implementing client level locking in separate table. Am I thinking in right direction. Please suggest. |
#3
| |||
| |||
|
|
Hello pm, AFAIK the # and ## are not memory tables they are tables in tempDB with differing visibility. The issue you will face even before getting there though is to do with Excel. Are people going to get the same Excel file?, dump to the same Excel File? Allan My process includes following processes for various clients. - loading excel file into a sql database - Processing loaded data in sql tables - Dumping excel reports I implemented these three stages by using set of physical temporary sql tables (not the memory ones with # and ##) for data transformations etc. I used to delete the data in these temporary tables and reuse for the next client. Till now I was the only person taking care of complete thing. Now I want to add multi-user capability to the system. I feel the only way to implement this is to change all my physical temporary sql tables to memory temporary tables i.e. using (# ## tables). But sql statement in data transformation task does not takes # tables. So I will have to dynamically change the sql to have # tables. Is it ok to load 20k records into memory for each client at the same time. Or is it a good idea to use physical temporary tables i.e. seperate physical table for each client and delteting it at the end. Also implementing client level locking in separate table. Am I thinking in right direction. Please suggest. |
#4
| |||
| |||
|
|
Hello pm, AFAIK the # and ## are not memory tables they are tables in tempDB with differing visibility. The issue you will face even before getting there though is to do with Excel. Are people going to get the same Excel file?, dump to the same Excel File? Allan My process includes following processes for various clients. - loading excel file into a sql database - Processing loaded data in sql tables - Dumping excel reports I implemented these three stages by using set of physical temporary sql tables (not the memory ones with # and ##) for data transformations etc. I used to delete the data in these temporary tables and reuse for the next client. Till now I was the only person taking care of complete thing. Now I want to add multi-user capability to the system. I feel the only way to implement this is to change all my physical temporary sql tables to memory temporary tables i.e. using (# ## tables). But sql statement in data transformation task does not takes # tables. So I will have to dynamically change the sql to have # tables. Is it ok to load 20k records into memory for each client at the same time. Or is it a good idea to use physical temporary tables i.e. seperate physical table for each client and delteting it at the end. Also implementing client level locking in separate table. Am I thinking in right direction. Please suggest. |
![]() |
| Thread Tools | |
| Display Modes | |
| |