[SOLVED] Seach problem, old text is still found

After 'migration', I had some Documents with only <p>Missing!!!</p>, which I used to run a script to update the info, so now most of the documents are updated.

Strangly, search still list the documents if I search for 'Missing!!!'.

I have cleared and rebuilt the catalog.
The site runs on a local ubuntu machine, so no nothing in front.

Any suggestion on how to debug this?

PS: The catalog rebuild returned a few 'Syntax Warning: Invalid number of shared object groups', but did finish ( [waitress-0] Catalog Rebuilt )


I guess somewhere in the objects theres still the text "Missing!!!".
Do you by any chance have a pdf or something like that in the object?

No, it is just plain Documents.


Should have thought about it, but the text acually contains 'missing', without any !!!

Turns out search for 'Missing!!!' also finds the word 'missing' (the texts are about boats, so some went 'missing' (especially during the war).

I used collective.searchreplace and replaced <p>Missing!!!</p> with <p>Missssing!!!

so now I can find the few remaining.

You can check in portal_catalog for each indexed object the exact values of indexed values per object.

What happened here is that the search text Missing!!! was cleaned to missing.

An example would be this , where the search text is turned into admin and therefor the catalog returns Restricted Python injection .

If you'd search for Missing!!!!!!!!! it'll probably work too.


The combination of me using using 'Missing!!!' for the content (which I did to mark that we miss (lack) some information about this item (boat) and (that) so many also had the (body) text 'missing' was not obvious 'after a long day'.
Today, it is more obvious. The reason we miss/lack information is often related to the fact that the boat went missing (a lot did during WWII).

1 Like