How do I normalize data during an import?

[Johannes Buverud]

In our SCCM database, there is both "Hewlet Packard" and "HP" as manufacturer. The same inconsequence is then imported into Insight. This makes filtering/serching/reporting difficult.

Other Asset tools I have worked with have a normalization alias functionality where you can set what values should be normalized during an import. In this case it would maybe set "Hewlet Packard" = "HP", and then all objects would get "HP" as manufacturer during the import. 

Is it maybe posible to use the regex functionality to solve this?

Will Riada make a normlization alias functionlality for Insight?

[PJ Wysota]

Nope at all.

It is all about normalisation and standardisation data within sources.

But there might be the option to do that just AFTER imports.

Firstly - please refer to blog post here: https://blog.valiantys.com/en/expert-tips/insight-for-jira

1. I would use HP as a reference to "manufacturer" Object rather than text attribute (for future standardisation)
2. For imports (Only) data I would use another like "Manufacturer TXT" text attribute

3. Map the import data against text one
4. Afterward - using filter - pick up both HP and Hewlett Packard and, using bulk edit/update, map against one from "Manufacturer" object

Of course - with one time import it works better. :-) as you do it once, and afterwards you can delete TXT attribute. If we are talking about synchronisation of ext data source over and over again, you would need to repeat this operation as well.

[Johannes Buverud]

I'll agree that this should be normalized within SCCM, but I think this is HPs own fault using different values in their config. 

I allready have Manufacturer (and also Model) as object types, being mapped when importing, Also, I have an alias attribute on Manufacturer allowing me to do aliasing when importing from different sources. But this does not work when the same source delivers inconsequent data.

We have a lot of data being imported, and they have to be correct, so any manual updating is not an option.

I could of coure make a script and use automation to trigger on create and update, this script had to be maintained, and that is also not a very flexible solution.

I believe aliasing definately should be a part of Insight out-of-box. When importing from different sources, there is always a lot of different manufacture names being used. (The alias list from our previous CMDB is quite long.) The same applies for Model.

[PJ Wysota]

Well, it should be. But I would rather say it is Feature Request for RIada (and I assume it would be for automations running rather on saved data, than on imports) - like Daily standardisation rule or sag.

[Danyal Iqbal]

Is it maybe posible to use the regex functionality to solve this?

Unfortunately No.

[Johannes Buverud]

Regarding regex, I was thinking of this excplicit case, namely changing the data locator with the value "Hewlet Packard" to "HP" before saving it to the attribute. Using the "Object type mapping"->"Configure row"->"Regular expression" 

I thought regex had a replace functionality? 

[PJ Wysota]

Rather filtering, and cutting off pre- and suffixes. You could e.g. get user names as text from LDAP/AD ou=, cn=, and similar strings.

[Johannes Buverud]

I did not understand your answer. What does user names in LDAP has to do with Manufacturer names in SCCM?

[PJ Wysota]

Nothing. But REGEX capabilities do. And what I explain is that REGEX will have no normalization capabilities here - but rather filtering data input.