Exact match when searching for custom field

[cimnine]

Issue type

[X] Feature request
[ ] Bug report
[ ] Documentation

Environment

• Python version: 2.7
• NetBox version: 2.2.4

Description

• Create a custom field:
  • Objects: dcim > rack
  • Type: Text
  • Name: test1
  • Is filterable: ✔️
• Take two racks and set the value of the custom field as follows:
  • Rack 1: urn:bla:1
  • Rack 2: urn:bla:100
• Go to the rack overview and put a urn:bla:1 into the filter for the custom field

Expectation: I get one result.
Current behaviour: I get two results.

Suggestions

I suggest that all filtering on fields become exact and at the same time introducing a wildcard character for fuzzy searches, e.g. *. (E.g. urn:bla:1 would only match the one rack, whereas urn:bla:1* would still match both racks.)

An alternative suggestion is to add an option to the custom field's definition: Shall that field match exactly or not. (Similar to the "Is filterable" checkbox.)

Another alternative suggestion is to support the Google syntax for exact matches, i.e. everything that's wrapped in " must be an exact match. (E.g. "urn:bla:1" would only match the one rack, whereas urn:bla:1 would still match both racks.)

[cimnine]

Some context

We're using custom fields to represent the IDs of our legacy system, which happen to be URNs. (These are all over the place, so we have to stick with them for a long time.) Currently, I'm writing a tool that copies any rack from the legacy system to Netbox and updates all racks already existing in Netbox based on the custom field. Hence I would appreciate if I could have a way to do exact-match lookups on custom fields.

[troxil]

@cimnine Isn't this how all existing filtering works? It's not exact at all. Prefixes, devices etc.

[cimnine]

Thanks for pointing this out, @troxil. I've changed it from a bug report to feature request.

[jeremystretch]

An alternative suggestion is to add an option to the custom field's definition: Shall that field match exactly or not. (Similar to the "Is filterable" checkbox.)

This seems like a favorable approach over introducing a potentially ambiguous wildcard approach.