Testing DQ checks: miscellany - Dates

[Mesibov]

Tested on 457075 beetle records downloaded 2 December 2015. It's difficult to test these with the download, which is missing a field for date supplied. Even so, some problems are evident.

Invalid collection date ('Unparseable date supplied from the original database')

444822 false
12253 true

Incomplete collection date ('The date supplied was incomplete, missing the day and/or month component')

413203 false
43872 true

There are 4 possible cases:

(1) invalid FALSE, incomplete FALSE
(2) invalid FALSE, incomplete TRUE
(3) invalid TRUE, incomplete FALSE
(4) invalid TRUE, incomplete TRUE

(1) These records should have an 'Event Date - parsed', unless no date details at all were supplied. There are, however, 88 false/false records with 'Year - parsed' and 'Month - parsed' fields occupied, but no 'Event Date - parsed'. Examining just one of them:

http:https://biocache.ala.org.au/occurrences/823f1736-73d9-4122-97bf-768b53331bb5

shows that the original record has the Event Date 1976-12-13, but the record has been processed as

Collecting Date
[date not supplied] Year: 1976, Month: 12, Day:
Supplied date "1976-12-13"

(2), (3), (4) None of these records should have an 'Event Date - parsed', but 3 do. Each has a mistaken 'Year - parsed' entry:

RecordID Year - parsed Month - parsed Event Date - parsed Incomplete Invalid
bbf7b7c0-b65f-4386-b1b6-530187359667 1279 01 1279-01-01 false true
a480ce45-ca53-4657-8c5d-1b44280ca586 0196 01 0196-01-01 false true
99b7100a-bcca-421f-9e7b-a3a2f6b7c5ee 0905 10 0905-10-01 false true

And this Queensland Museum record has an 'Event Date - parsed' of 1668-02-07:

http:https://biocache.ala.org.au/occurrences/97cc70f0-86c1-4833-a992-79d6250159a4

It was flagged 'false' under 'Invalid collection date'. QM says it's a mistake for '1968'.