opendatacz / VVZ_extractor

VVZ extractor DPU for ODCleanStore2
0 stars 0 forks source link

Why are there notices with more than 1 publication date? #5

Open jindrichmynarz opened 10 years ago

jindrichmynarz commented 10 years ago

Some contract notices have more than 1 publication date. See the results of this query:

PREFIX pc: <http://purl.org/procurement/public-contracts#>

SELECT ?contract ?notice
(COUNT(DISTINCT ?publicationDate) AS ?count)
(GROUP_CONCAT(?publicationDate; separator = ", ") AS ?tenderDeadlines)
WHERE {
  GRAPH <http://linked.opendata.cz/resource/dataset/vestnikverejnychzakazek.cz> {
    ?notice pc:publicationDate ?publicationDate .
    OPTIONAL {
      ?contract pc:publicNotice ?notice .
    }
  }
}
GROUP BY ?notice ?contract
HAVING (COUNT(DISTINCT ?publicationDate) > 1)
ORDER BY DESC(?count)

Again, it might be relevant to #3 and the true cause of the issue might be inadvertent merge of contracts. Some of the notices that have multiple publication dates are related to contracts that suffer from #3.

michalpomykacz commented 10 years ago

Contract may be represented by multiple forms (regular, and then corrections). For example see http://www.vestnikverejnychzakazek.cz/cs/Searching/RelatedContracts?Code=372320, where under the same form number is two same contracts differing by publication date

jindrichmynarz commented 10 years ago

These are the publication dates of public notices, not contracts. Why is it the case that the same public notice (i.e. identified with the same URI) can be published on multiple dates? I think this is either error in the original data, error in extraction, or it the data should be split into 2 public notices.