CA Service Management

Hi team,

what is the best approach to use long numbers in CMDB?

AFAIK there are three possibilities:

1. use integer     It is OK until the number are smaller than the biggest int - then it overflows and is displayed as negative number (and this is basically what happened in our implementation)

2. use string       Using string is fine until you need to sort the results by this attribute - users are confused and complain (and this is what we did as a first solution)

3. use double     Double is almost perfect until you need to use grloader - and we need to use it for regular data synchronization which can't be done by pdm_userload to usp_owned_resource table

4. use something I don't know

Thanks for any ideas,

Jakub

Jakub, I think option 2. use string should be the one. The end users

complain and are confused because you don't format the string value

properly. I think you can(for example using some js script) format

the user input for this value  in such a way that it has to be entered as integers and the integer will be formatted before saved. For example,

123 will be 0000000123 and 23 will be 0000000023 and therefore

later when sorted 23/0000000023 will be before 123/
0000000123. Thanks _Chi

Hi Chi,

thanks for your idea.

I should be able to save number 123 as 000000123 using JS or trigger+spel and I agree this can solve the sorting issue. But think about the searching - users will definitely search for '123' and not for '000000123'.

Jakub

Jakub, I agreed this is not the perfect one but I do think this should be better than other options. As for the "search" part, we may think this as user education thing and/or you give some tip on the search page(list_xxxx.htmpl) so the end user know

he/she should enter "%123" to search for "123". Thanks _Chi

Hi Chi,

using of '%123' has another confusing side-effect - it returns also 654619123 and 78911123.

I still hope a bit that this is more about my lack of knowledge and not about the solution itself. If there is no other option I will use double (which is transparent from the user point of view) with some additional web services scripts for importing these attributes.

Thanks,

Jakub

In the end I created completely different solution which combines string and double attributes. I simply use the string attribute as the input for grloader. Changing the string attribute triggers simple spel which changes the double attribute used in GUI. The synchro works in both ways - if you change the double attribute in GUI, another spel changes the string attribute to the same value.

I haven't found any side-effects or performance issues.

Regards,

Jakub