See https://jira.bikalabs.com/browse/LIMS-1988 re dealing with multiple instruments using the same keywords.
To get pass it at Metron, we can use the full LineName from the file, e.g. Ca 315.887 but strip the space and . out of it as keywords do not allowed that in the LIMS. 315.887 refers to the wavelength in nm of the Ca's spectral analysis.
e.g. read Ca 315.887 in the file and look up Ca315887 in the system, Mg 279.806 for Mg279806 etc. The element abbreviation and wavelength combination is pretty unique the lab ensures me.
Please quote for this customisation. No doubt we'll run into this eslewhere too where the line name cannot be used and then we can reconsider the deluxe solution at https://jira.bikalabs.com/browse/LIMS-1988.
More from IRC:
<lemoene> rockfruit, do you know whether AS keywords will handle spaces and full stops, e.g 'Ca 315.887' ?
<rockfruit> spaces? I really want to say no to spaces. let me check.
<lemoene> maybe we could just strip them out. NP
<lemoene> full stops?
<lemoene> Ca 315.887 is the wavelength the ICP test for and appears like that in the results file
<rockfruit> accepts [A-Za-z\w\d\-\_] - that's a-z, A-Z, word-characters, numbers, dashes and underscores. anything else fails validation
<lemoene> just strip them out before loading and keep a keyword Ca315887 in the LIMS?
<rockfruit> I'm not sure why the validator is so restrictive. offhand, can't think of a reason we couldn't allow keywords to be anything they want to be. BUT I'm sure we had a reason.... so needs some exploratory digging.
<lemoene> if stripping them out works, fine
<rockfruit> We use it for the ID of analyses in the AR... there are some restrictions there..
<lemoene> ok. not lose time
<rockfruit> yeah stripping sounds good, and is also a good plan. stripped, then the same no-duplications check applied. cannot clash with another service keyword, and cannot clash with a calculation interim field's ID.