Getting rid of orphan article elements after articles were purged

Build 1501 on 14/Nov/2017  This topic last edited on: 24/Oct/2016, at 13:00

In some situations, orphan article elements can remain in database after articles are purged (as explained in Photocaptions survive article deletion), and prevent purging of linked images.

You may want to get rid of such orphan elements, but it is better not to use SQL queries, as elements data is spread across many tables and incomplete deletions can leave "dirty" data.

This topic explains step-by-step how to build a XPath expression to clean such orphans, both article elements and referenced images.

Let's start with an article that has a photoCaption element and a linked image. Open the article in Ted4, open the photoCaption element and then display its properties. On the Info tab find the ID of the photoCaption article element. Let's assume the ID is 130031.

Click to toggle graphic sizeBuild XPath by looking into a regular photoCaption

Open the browser and sign in to GN4.

Duplicate the window, and then paste in the address bar the following:

http://main.teradp.com/GN4/do.ashx?cmd=objsNav&ids=130031&options=referencing&max=6

The display will be something similar to this:

Click to toggle graphic size

If you look at the <nav:refObject idref="obj129847"> node you'll find the id of the parent article. Therefore, the XPath that connects a photocaption with its article is:

gn4:photoCaption/nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject

How we got it?

Let's cut out of the above XML all parts that are not important for this task, and then let's examine what remained, starting from nav:referencing - you see the sequence of nodes specified above in the XPath:

<photoCaption nav:objectType="photoCaption">

 <nav:objectTypeAttrs ...></nav:objectTypeAttrs>

         <nav:referencing>

                 <nav:parentBy attrName="article.txts">

                         <nav:refs handle="1">

                                 <nav:refObject idref="obj129847">

                                 </nav:refObject>

                         </nav:refs>

                 </nav:parentBy>

Now, this handles the photoCaption. What about other orphans?

The XPath that connects any article element with its article is:

gn4:txt/nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject

However, we're looking for article elements that do not have a parent article, therefore, let's "negate" the condition with a not enclosing all:

gn4:txt[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]

However, we have to think also on images, linked to orphan photocaptions. The XPath to find such items is more complicated:

gn4:img[nav:referencing/nav:referencingBy[@attrName="photoCaption.ref"]/nav:refs/nav:refObject/gn4:photoCaption[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]]

Let's summarize:

Xpath to locate all orphan photoCaption elements

gn4:photoCaption[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]

Xpath to locate all orphan article elements of any type

gn4:txt[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]

Xpath to locate all images linked to orphan photoCaption elements

gn4:img[nav:referencing/nav:referencingBy[@attrName="photoCaption.ref"]/nav:refs/nav:refObject/gn4:photoCaption[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]]

Note: The complete list of navigation options:

http://tech.teradp.com/tech/html/gn4/docs/VSdoc/index.html#frlrfTeraDPGN4GNClientDataNavigatorOptionClassTopic.html

Click to toggle graphic sizeUse a script to get ids of all orphan elements

Now, one can write a small VB.Net script to generate the list of ids of the orphan elements, and run it from the Script editor in Ted4 or Fred4:

Click to toggle graphic size

    'find the ids of orphan txts

    'get the login

    Dim clientLogin As TeraDP.GN4.Client.ClientLogin = DataConnection.Instance.ClientLogin

    'create the GNQuery conditions

    Dim sc As New ExtSearchConditions("gn4:txt[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]")

    'execute the query:

    'the second parameter True is to return the short name instead than the full descriptive name

    'SpikedHandling.Exclude is to ignore the spiked objects

    Dim idsList As System.Collections.Generic.List(Of IdName) = clientLogin.ListObjs(sc, True, SpikedHandling.Exclude)

    'create a comma-separated string containing the ids

    Dim idsString As String = Nothing

    If idsList IsNot Nothing Then

      For Each idName As IdName In idsList

        idsString &= CStr(idName.Id) & ","

      Next

    End If

    'save the result string on file

    If Not String.IsNullOrEmpty(idsString) Then

      System.IO.File.WriteAllText("c:\temp\orphanTxtIds.txt", idsString)

      MsgBox(String.Format("Found {0} orphan texts. Saved on c:\temp\orphanTxtIds.txt.", idsList.Count))

    Else

      MsgBox("No orphan text found")

    End If
Click to toggle graphic sizeSave the content of orphan items before deleting them

If you want to save the content of orphan items before deleting them, use the following cmd4 command (all in one line):

cmd4 get 

 -out c:\temp\orphans.xml

 -conditions "gn4:txt[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]" 

 -srvurl http://myhost.com/GN4/Adminsrv.asmx 

 -UserName xxxx 

 -password yyyy

Note about the -srvurl parameter

In version 2.5 and newer,  the URL is http://<myserver>/<myWebApp>/. In version prior to 2.5, the URL is the address of the Web service to connect to, e.g. http://<myserver>/<myWebApp>/Adminsrv.asmx.

The result may be:

C:\tera\GN4\main\bin>cmd4 get -conditions "gn4:txt[not(nav:referencing/nav:paren

tBy[@attrName='article.txts']/nav:refs/nav:refObject)]" -username xxxx -passwor

d yyyy -out c:\temp\exportedorphans.xml -srvurl http://main.teradp.com/gn4/admin

srv.asmx

Cmd4 version 2.1.2084.0

GN4 command-line client

Copyright © 2006 - 2014 Tera DP srl

Language en-US

 

Getting object(s)...

Saving object(s) to file 'c:\temp\exportedorphans.xml'...

The exported content looks like:

Click to toggle graphic size
Click to toggle graphic sizeSpike orphan items by srv4

When you want to spike orphan items, you need to do it in a correct order: first spike the images, linked to the orphan photocaptions and their geometries; then, spike the other orphan elements, if any.

Therefore, first execute the command to spike images and geometries, linked to orphan photocaptions(all in one line!):

srv4 spike 

 -conditions "gn4:img[nav:referencing/nav:referencingBy[@attrName='photoCaption.ref']

 /nav:refs/nav:refObject/gn4:photoCaption[not(nav:referencing/nav:parentBy[@attrName='article.txts']

 /nav:refs/nav:refObject)]]" -extend photoCaption.ref,txtGeometry.txtRef –recurse

Then, execute the command to spike other orphan elements (all in one line!):

srv4 spike 

-conditions "gn4:txt[not(nav:referencing/nav:parentBy[@attrName='article.txts']/nav:refs/nav:refObject)]" 

-extend txtGeometry.txtRef