"Page URL Does not contain" segment condition has excluded Page URL in reports

[Starker3]

It looks as though the "Page URL does not contain" segment condition still allows some Page URLs to be included in the Pages report.

If we take an example of this segment:

And then process the reports, we still get some Page URLs in the report that contain this exact string:

However, the segmented visits log doesn't show any actual data for this:

When searching the actual visits log there are no visits that include this string:

[NathanDB99]

I have brought this issue to the attention of Matomo and they were able to replicate is, as seen above. In my case the /index of pages we want to exclude are still shown in the pages report. Indeed the /index of this pages are not shown in the visitor logs.

[sgiehl]

I think that is kind of expected behavior as segments in Matomo work that way. A segment is always used to filter the list of visits that are included in a segment. So when using a segment like page url does not contain Matomo will select all visits that have a page view with a url that does not contain something. It will not exclude a visit if there is an action that contains the defined url (if there are also other actions). Afterwards when generating the reports the excluded url will so be contained in the reports again, as it is part of a visit that is included in the segment.

[Starker3]

@sgiehl But in that case would we not expect to see that in the Visits Log when viewing the segment?

[heurteph-ei]

Isn't is the same as: https://github.com/matomo-org/matomo/issues/17388 ❓

[sgiehl]

@Starker3 you will also see that in the visits log. What you wouldn't see there is a visit that only has an action with the excluded url. But as soon as visit also has other actions it will be shown.

[tsteur]

FYI And it likely refs https://github.com/matomo-org/matomo/issues/14664

[Starker3]

Thanks for the additional info @sgiehl When I check the Visits log for All Visits we can indeed see some visits that contain other actions:

However, if I apply the segment from the screenshot above (That excludes any Page URL with /selection), then there are no visits shown that include this Page URL in all ~900 visits in the segment:

[9joshua]

A customer has a related issued with segment filters where multiple "Does not contain" filters are used.

If only one filter is used (e.g. Page URL Does not Contain '888'), any visit with at least one action matching that condition causes the entire visit to be excluded, as expected...

However, when multiple "Does not Contain" conditions are joined by an OR operator, the visits where BOTH conditions are contained in the SAME URL are excluded... The only excluded URL above was localhost/?888&999

Expected behaviour: When excluding a URL via "Does not Contain" one would expect the filter to omit the entire visit where the condition is matched. Similar to how a "Contains" condition would include the entire visit where the condition is met. So a visit that included both of the defined URLs should be excluded.

Actual behaviour (with multiple "Does Not Contain" conditions are joined with OR): IF:

• Any page URL in the visit does not contain 888, OR
• Any page URL does not contain 999, THEN
• The visit IS included in the segment.

This makes it impossible to filter a report for visits where a visitor did not visit two specific URLs that match the "Does not Contain" conditions.