morgandawn: (Cat Sleepy)
[personal profile] morgandawn
 If you are designing an Excel database with creator names how do you handle entries where the creator is listed as "possibly Morgan Dawn"? 

I can think of several ways...

Creator:  Unknown-Morgan Dawn might be best.  That way if you sort, all the Unknown-Morgan Dawn works will still be listed together.  They won't be listed next to the confirmed Morgan Dawn works, but still searchable. If and when the ID is confirmed it could then be moved to "Morgan Dawn" or even updated to "Mary Sue Lamb". It could also handle multiple guesses: "Unknown-Morgan Dawn-Mary Sue Lamb

The idea is to leave "Unknown" for those were we have no clue, and keep the ones where we are guessing in another grouping.

The other way is to have a second column that would add a Y/N in  "Presumed Creator" but I don't like separating data. Plus one more cell to fill in is a pain.

(no subject)

Date: 2017-09-17 04:31 am (UTC)
laisserais: kiss (Default)
From: [personal profile] laisserais
The best way to maintain data integrity is to separate the known/unknown into a separate column. The reason is that the concept of known/unknown is not the same as the concept of a writer's proper name, and therefore needs to be a separate field.

Especially since you want to use unknown as a choice for name, moving their status to a separate flag column will give you maximum flexibility for the future, as well as aggregability.

(no subject)

Date: 2017-09-17 04:33 am (UTC)
laisserais: kiss (Default)
From: [personal profile] laisserais
oh and as for it being a pain: yes, but you can get around this if you have already started your project. You can insert columns, then choose 'text to columns' on the data ribbon. Choose 'delimited' and then tell it to separate on the dash or whatever you've used; it'll do a mass edit on the whole column. :)

(no subject)

Date: 2017-09-17 05:05 am (UTC)
laisserais: kiss (Default)
From: [personal profile] laisserais
So your field headers would be:

Vidder Name | Confirmed (y/n) | Name of Vid | etc

and the data would populate like so (each return is a new row):

Morgan Dawn | Y | your-vid-name | etc
Unknown | N | another-vid-name | etc
Mary Sue Lamb | N | another-vid-name | etc

and you're probably not gonna like this, but to handle multiple guesses, the best way to do it is to create a new row for each guess, with only the vidder name being different.

That's not intuitive, because it feels like it's redundant--and if you're only ever gonna be using this for personal use and only reading it with your eyes, then you don't have to--but if this is ever going to grow or be used as an actual database, then it's best to do it that way. I call it 'machine readable' -- it's called first normal form.

In an actual database, you'd want to do this and attach a unique identifying number to every vid name and every vidder name.

Doing it like that will also help you set up pivot tables, where you can condense info and sort quickly.

(no subject)

Date: 2017-09-17 07:30 am (UTC)
laisserais: kiss (Default)
From: [personal profile] laisserais
Here's an example of what I mean. If you want to build something that will be useful long term, this is the way to go. I think you're on the right track!


morgandawn: (Default)

October 2017

1 2345 6 7
8 9 10 11121314

Most Popular Tags

Page Summary

Style Credit

Expand Cut Tags

No cut tags