ess-dive-community / essdive-file-level-metadata

READY TO USE. Reporting format for File Level Metadata uploaded to the ESS-DIVE repository
https://ess-dive.gitbook.io/file-level-metadata-reporting-format/
Creative Commons Attribution 4.0 International
1 stars 3 forks source link

missing value codes #17

Closed tvelliquette closed 3 years ago

tvelliquette commented 3 years ago

I suggest the following changes:

Possible issues with potential solution

  1. I suppose if the file doesn’t contain any missing values then inserting “N/A” doesn’t actually change anything since that value doesn’t exist in the file.
  2. If there are empty cells with no missing values used to represent those empty cells that brings up a different need. Should that be represented as “blank” or “ ” or “missing values are represented by blank cells”
  3. Recommend explain in greater detail about where and how the missing value codes are assigned and what they represent in the Data Dictionary Description field(s). You could add the same or more brief description in the FLMD Notes field.

Need to refer users to the CSV Detailed Guide under Delimiter for how to format fields with multiple entries that contain commas.

robcrystalornelas commented 3 years ago

I made a related issue under the CSV reporting format repo, since that is where we have guidance on not using semicolon within cells: https://github.com/ess-dive-community/essdive-csv-structure/issues/6

tvelliquette commented 3 years ago

We have since decided that a "semicolon" is okay and that wording needs to be changed in the documentation.

tvelliquette commented 3 years ago

I can't seem to get the edit I made to the FLMD quick guide to display. I see it in Markdown.

I added instructional text about using delimiters for multiple missing value code entries rather than forcing users to go back to the CSV delimiter instructions. If you think the text is good and you can get it to commit changes, close it out.

Here is what I tried to commit: Report all Missing Value Codes. For columns containing numeric data, use "-9999" as the missing value code (or modify to match significant figures given the data). For columns containing character data, use "N/A" as the missing value code. If entering more than one missing value code, use a vertical bar "|" or semicolon ";" instead of a comma or protect the comma with matching double quotation marks around the entire value. For more information about commas not meant to be a delimiter (e.g. used within a cell), refer to the Delimiter section of the CSV Reporting Format Detailed Guide.

robcrystalornelas commented 3 years ago

All set now. I copied and pasted over the new text you provided above to the missing value section of the quick guide.