10

Please login or register to vote for this query.

(click on this box to dismiss)

This query lists questions that have been closed as duplicates of another question that has since been deleted. Such questions should be either deleted or, if they have good answers or are otherwise of decent quality, reopened. Note that non-locked questions with no answers and a negative score will be automatically deleted by the RemoveDeadQuestions script after 30 days, as indicated by the "Roombable?" column. By default, the results will be sorted with the most recently closed questions first. You can click the column headings to change the sort order. For example, questions with a high answer score may be good candidates for reopening (or for a historical lock), while questions with a high score but no answers may simply need a new dupe target (see http://meta.stackexchange.com/q/291824).

Q&A for people interested in the history and origins of science and mathematics

-- step 1: find IDs of dupes with deleted targets
CREATE TABLE #DeletedDupes (
  PostId INT NOT NULL,
  RelatedPostId INT NOT NULL
)
INSERT INTO #DeletedDupes
SELECT
  BrokenLink.PostId,
  BrokenLink.RelatedPostId
FROM
  PostLinks as BrokenLink
  LEFT JOIN Posts as TargetPost
    ON TargetPost.Id = BrokenLink.RelatedPostId
WHERE
  BrokenLink.LinkTypeId = 3
  AND TargetPost.Id IS NULL
  AND NOT EXISTS ( 
    SELECT 1 FROM
      PostLinks as OtherLink
      JOIN Posts as OtherTarget
        ON OtherTarget.Id = OtherLink.RelatedPostId
    WHERE
      OtherLink.LinkTypeId = 3
      AND OtherLink.PostId = BrokenLink.PostId
  )
  
-- step 2: determine current site host name for migration matching
DECLARE @sitename NVARCHAR(MAX) = DB_NAME()
SET @sitename = REPLACE(@sitename, 'StackExchange.Mathoverflow', 'Mathoverflow')
SET @sitename = REPLACE(@sitename, 'StackExchange.Ubuntu', 'AskUbuntu')
SET @sitename = REPLACE(@sitename, 'StackExchange.Audio', 'StackExchange.Video')
SET @sitename = REPLACE(@sitename, 'StackExchange.Garage', 'StackExchange.Mechanics')
SET @sitename = REPLACE(@sitename, 'StackExchange.Moderators', 'StackExchange.CommunityBuilding')
SET @sitename = REPLACE(@sitename, 'StackExchange.Photography', 'StackExchange.Photo')
SET @sitename = REPLACE(@sitename, 'StackExchange.Programmers', 'StackExchange.SoftwareEngineering')
SET @sitename = REPLACE(@sitename, 'StackOverflow.Br', 'StackOverflow.Pt')

DECLARE @tld NVARCHAR(3) = 
  (CASE WHEN @sitename LIKE 'Mathoverflow%' THEN 'net' ELSE 'com' END)

DECLARE @newurlprefix NVARCHAR(MAX) = 'https://' + LOWER(CASE
  WHEN CHARINDEX('.', @sitename, 0) = 0 THEN @sitename
  ELSE RIGHT(@sitename, LEN(@sitename) - CHARINDEX('.', @sitename, 0))
    + '.' + LEFT(@sitename, CHARINDEX('.', @sitename, 0) - 1)
END) + '.' + @tld

DECLARE @oldurlprefix NVARCHAR(MAX) = ''
SELECT @oldurlprefix = CONCAT(value, '.', @oldurlprefix)
  FROM STRING_SPLIT(@sitename, '.')
SET @oldurlprefix = 'http://' + LOWER(@oldurlprefix) + @tld

CREATE TABLE #UrlPrefixes (
  UrlPrefix NVARCHAR(MAX) NOT NULL
)
INSERT INTO #UrlPrefixes VALUES (@newurlprefix), (@oldurlprefix)

-- step 2.5: compile list of migration record patterns to match
CREATE TABLE #MigrationPatterns (
  PostId INT NOT NULL,
  Pattern NVARCHAR(MAX) COLLATE SQL_Latin1_General_CP1_CS_AS NOT NULL
)
INSERT INTO #MigrationPatterns
SELECT PostId,
  'from ' + UrlPrefix + '/questions/' + Cast(RelatedPostId as NVARCHAR) + '/%'
FROM #DeletedDupes, #UrlPrefixes

-- step 3: filter out dupes of migrated posts
CREATE TABLE #MigratedDupes (
  PostId INT NOT NULL
)
DECLARE sites CURSOR FOR
  SELECT name
  FROM sys.databases
  WHERE name NOT IN ('master', 'tempdb', 'model', 'msdb', 'Data.StackExchange')

DECLARE @sitedbname AS nvarchar(max)
DECLARE @query AS nvarchar(max)

OPEN sites
FETCH NEXT FROM sites INTO @sitedbname
WHILE @@FETCH_STATUS = 0
BEGIN
  SET @query = '
    USE [' + @sitedbname + ']

    INSERT INTO #MigratedDupes
    SELECT
      Match.PostId
    FROM #MigrationPatterns as Match
      JOIN PostHistory as Migration
        ON Migration.Text LIKE Match.Pattern
    WHERE
      Migration.PostHistoryTypeId IN (17, 36)
    '
  EXEC sp_executesql @query
  FETCH NEXT FROM sites INTO @sitedbname
END
CLOSE sites
DEALLOCATE sites

-- step 4: merge and uniquify results, add aggregate columns
CREATE TABLE #FilteredDupes (
  PostId INT NOT NULL,
  AnswerScore INT NULL,
  LockEventId INT NULL
)
INSERT INTO #FilteredDupes
SELECT
  BrokenLink.PostId as PostId,
  MAX(Answer.Score) as AnswerScore,
  MAX(LockEvent.Id) as LockEventId
FROM #DeletedDupes as BrokenLink
  LEFT JOIN Posts as Answer
    ON Answer.PostTypeId = 2
    AND Answer.ParentId = BrokenLink.PostId
  LEFT JOIN PostHistory as LockEvent
    ON LockEvent.PostHistoryTypeId IN (14, 15)
    AND LockEvent.PostId = BrokenLink.PostId
WHERE
  BrokenLink.PostId NOT IN (SELECT PostId FROM #MigratedDupes)
GROUP BY
  BrokenLink.PostId

-- step 5: format results nicely
SELECT
  BrokenLink.PostId as [Post Link],
  SourcePost.ViewCount as Views,
  SourcePost.Score,
  SourcePost.AnswerCount as Answers,
  BrokenLink.AnswerScore as AScore,
  (CASE WHEN SourcePost.AcceptedAnswerId IS NULL THEN '' ELSE 'Accepted' END) as Accepted,
  SourcePost.ClosedDate as Closed,
  (CASE
    WHEN LockEventId IS NOT NULL AND (SELECT PostHistoryTypeId FROM PostHistory WHERE Id = LockEventId) = 14 THEN 'No (Post Locked)'
    WHEN SourcePost.AnswerCount = 1 THEN 'No (1 answer)'
    WHEN SourcePost.AnswerCount > 1 THEN 'No (' + Cast(SourcePost.AnswerCount as VARCHAR) + ' answers)'
    WHEN SourcePost.Score = 0 THEN 'If downvoted'
    WHEN SourcePost.Score > 0 THEN 'If downvoted x' + Cast(SourcePost.Score+1 as VARCHAR)
    ELSE 'Yes, on ' + Convert(VARCHAR(10), DateAdd(Day, 30, SourcePost.CreationDate), 126)
  END) as [Roombable?],
  SourcePost.Tags as [Tags]
FROM
  #FilteredDupes as BrokenLink
  JOIN Posts as SourcePost
    ON SourcePost.Id = BrokenLink.PostId
ORDER BY
  SourcePost.ClosedDate DESC

view current version

edited dec 9 17 Ilmari Karonen

created feb 19 14 animuson