Ghost Search Results

[PHK]

Is it possible to eliminate individual highlighted areas that are produced by a Search but have nothing apparent to do with the actual search terms? Below is a screenshot of an area of a page where a search was conducted. In this screenshot, one can see highlighted areas that were produced in a conventional search. The four yellow rectangles are what I would call. "Ghost Search Results."

image.png

These appear in additon to legitimate search results. I have no idea how they are created.

[Daniel - PDF-XChange]

Hello, PHK

Is there any chance that there is invisible text content there? The Search function should not be able to place these results highlights unless there is a term in that location, so I suspect, if you use the "select text" tool, and drag a box over that area, you will find there is infact transparent or obscured text in that location.

Kind regards,

[PHK]

Thanks for thinking about this, Daniel, but this is what I get if I do as you suggest.

image.png

[Daniel - PDF-XChange]

Hello, PHK

That definitely indicates that, as suspected, there is text present there, and so that is why you get results in that location.

Kind regards,

[Willy Van Nuffel]

So, when you look into the Content-pane or the Comments-pane for the corresponding page, there must be some text equal to the term(s) in the Search pane?

Can it be that you are using a specific color theme or "Override document colors" in Accessibility Options?

Eventually post an extracted page of the PDF in question, in case it does not contain sensitive information.

Kind regards.

[PHK]

Hello, PHK

That definitely indicates that, as suspected, there is text present there, and so that is why you get results in that location.

Kind regards,

How do I visualize that text if it is there?

[PHK]

So, when you look into the Content-pane or the Comments-pane for the corresponding page, there must be some text equal to the term(s) in the Search pane?

This page is really a sheet of many (<100) merged pages with many legitimate occurrences of the text of the search terms, so that is not practical.

Can it be that you are using a specific color theme or "Override document colors" in Accessibility Options?

No.

Eventually post an extracted page of the PDF in question, in case it does not contain sensitive information.

Thanks, but I won't be doing that as the sheet contains ten years of bank statements.

Thanks for thinking about it though, Willy.

[rakunavi]

Hello PHK,

How do I visualize that text if it is there?

Open the Content Pane and Properties Pane in advance, select the text from the Options menu in the Content Pane, and change the Fill Color to the appropriate color to make the transparent text visible.

• Animation.gif

Best regards,
rakunavi

[PHK]

Thanks, rakunavi, but there is no text to color. Besides, that would apply to all search results, some of which are legitimate results.

One other piece of information that I neglected to pass on: I have used the Split Pages occasionally to get to this large sheet of many pages. I am beginning to think that is the problem. Something is going on when I Split Pages that messes up the Content. I think each of these Search hits were before Splitting Pages legitimate results selecting the text in the right place. The Search results is about the right count, about what I would expect, but these Ghost boxes are no longer associating with where the text is on the page. I conclude that something gets scrambled/lost when I Split Pages. Another sheet in the same file that has not had any pages split produces 'correct' results. Either I am torturing the product or there is a bug.

[Daniel - PDF-XChange]

Hello, PHK

I am sorry, but your screenshot from using the Select text tool proves otherwise. Did you mimic the actions in Rakunavi's Gif? You should be able to directly select all Base content text, which is what the select text tool was highlighting when you dragged a box over the area.

You do not need to change the text color to a vibrant color like Rakunavi did, you can simply use black or whatever color the remainder of the text in your document might happen to be. Alternatively, you can use the "Edit > text" tool to select the text in that section of the page (via marquee drag select) to ensure that you are only selecting the relevant text.

As for the possibility of split pages being related, While that sounds unlikely, I cannot rule it out. Unfortunately the only way to look into it would require you to send us a copy of the file, so, if thes ever happens to appear in a document which contains less sensitive information, can you please email it to us, with a brief description and a link to this forum post, so that we can take a look? (ideally, if you can explain how to reproduce it that would be great!)

Kind regards,

[PHK]

...
I am sorry, but your screenshot from using the Select text tool proves otherwise.

Oh, yeah? What then does it prove?

Did you mimic the actions in Rakunavi's Gif?

No, I didn't. But here is a screenshot of a typical area.

image.png

As you should be able to see, there are the blue bounding boxes for text. But also, there are these yellow Ghost boxes with no text of any color in them.

As for the possibility of split pages being related, While that sounds unlikely, I cannot rule it out.

I have just done an experiment that to me confirms that it is the splitting and re-merging pages that is the cause of the problem. I did a Search of a file that had a new sheet that had been made up of about 75 pages and that produced 84 hits of the Search term. I split that page into 4 sheets and conducted another Search. Again, 84 hits. Next, I re-merged those 4 sheets but I re-ordered, or scrambled, the pages. Now, the search produces 123 hits consisting of all the old "good" hits additonally with, presumbably, 39 "bad" hits.

Unfortunately the only way to look into it would require you to send us a copy of the file, so, if thes ever happens to appear in a document which contains less sensitive information, can you please email it to us, with a brief description and a link to this forum post, so that we can take a look? (ideally, if you can explain how to reproduce it that would be great!)

Yeah, maybe.

This discussion and my local experiments have made it quite clear to me that there is a problem with the text Content is coming out of splitting and remerging pages and I will be very careful about doing that in the future until, or if, you guys fix it. You probably won't want to waste any resources on fixing it until you've got more users complaining about it so I will not be holding my breath.

[Stefan - PDF-XChange]

Hello PHK,

There is an "XForm" object on the left side of your screenshot that seems to have been selected - so likely inside that XForm there is your invisible text.
It would really help us if you can share a copy of the page (you can e-mail it to support@pdf-xchange.com if sensitive) - so that we can take some screenshots on our side to illustrate how you can make only a specific occurence visible.

E.g. - please see the attached file

invisible text test.pdf

It has the word "text" three times in it - one of the lines is with a text that has "Fill colour" of none - and it would initially appear as if you have only two lines of text - but when you do a search for the word "text" in my sample - you get three highlights:

image.png

We do suspect the same happens with your file.
You can then switch to the "Edit Contents" tool - select your currently invisible text - and change the Fill colour of only that, leaving the other occurrences of the search term unchanged.

image(1).png

Of if you are certain this text is surplus - you can remove it from your file once it has been selected.

Kind regards,
Stefan

[PHK]

I fully understand and appreciate that Content text will not be visible if there is no Fill color.

These discussions of Content text miss my point altogether.

What I am going on and on about is Search results and what happens when I Split Pages and Merge Pages. The Search function is throwing up up results that are not related to any text in Content or on the visible page. How is that not clear?

Maybe later when I have time I will try to construct a non-sensitive example.

[Daniel - PDF-XChange]

Hello, PHK

After some additional thought on this, I recalled, the "split pages" offers some controls which are enabled by default and will result in the original content being retained as you describe. Can I ask what settings you have set here?

image.png

Also, could you please try changing your settings to what I show in my screenshot, and let me know if that helps?

Kind regards,

[PHK]

image(1).png

... Can I ask what settings you have set here?

The opposite of yours: the other size option and the box unchecked.

Also, could you please try changing your settings to what I show in my screenshot, and let me know if that helps?

Same unsatisfactory result.

Okay, okay, you have goaded me into providing a non-sensitive example, a product user manual.

image.png

First, Merge Pages. all pages.

image(1).png

Search this new sheet for "System".

image(4).png

Looks right, no?

Split Page all ways until you are back to the same basic pages.
Merge Pages but this time, check box Reverse Order.

image(9).png

Repeat Search.

image(2).png

Now the page is littered with Ghost results in addition to the correct results.

THIS is what I'm talkin' about.

[PHK]

Sorry, I neglected to attach the file. See https://1drv.ms/b/s!AsBVtdXZAI89kOdwVQ6t9le2eq2u1w?e=b4tIAt

[Daniel - PDF-XChange]

Hello, PHK

Thank you for the example file. This was not an attempt to goad you, but I genuinely could not offer any more help without seeing your settings and a sample file you have reproduced the issue in.

Now, I have processed your file and I note that when I did the split with the settings that I mentioned previously, specifically, choosing "Change the physical size of the page" and unchecking "preserve original content" the result had no extra "invisible" items present.

image.png

image(1).png

Please do ensure you are setting both of these options and you should not encounter any of the aforementioned issues with these files.
In effect what happens here is cropping. The Split pages tool actually duplicates the page into a number of duplicates that are required in the final output, and then crops each of them down to a set size with the respective content area therein.
When you "crop" a page, the original content can be fully removed, or it can be retained so that you can later choose to undo the crop. When the latter happens (as is the case with the Split pages tool's default settings), the original content is left in place and thus is included when you merge pages. However, as it is beyond the page bounds and has been hidden, it is not visible on the page either.
The Settings I showed will ensure that the former is the case, and all content past the edge of the pages is removed and you cannot "undo" the crop process, which in turn ensures that there is no hidden content applied during the merge.

Kind regards,

[PHK]

...
Please do ensure you are setting both of these options and you should not encounter any of the aforementioned issues with these files.
In effect what happens here is cropping. The Split pages tool actually duplicates the page into a number of duplicates that are required in the final output, and then crops each of them down to a set size with the respective content area therein.
When you "crop" a page, the original content can be fully removed, or it can be retained so that you can later choose to undo the crop. When the latter happens (as is the case with the Split pages tool's default settings), the original content is left in place and thus is included when you merge pages. However, as it is beyond the page bounds and has been hidden, it is not visible on the page either.
The Settings I showed will ensure that the former is the case, and all content past the edge of the pages is removed and you cannot "undo" the crop process, which in turn ensures that there is no hidden content applied during the merge...

Thank you, Daniel, for that advice and elaboration on what those options do. Frankly, the online user guide (excerpted herein below)

image.png

comes nowhere near explaining what these options do, in my humble opinion. I wish I had known what you just explained 18 months ago when I started to use more extensively the Merge Pages and Split Pages functionality. It would have saved me many, many hours and headaches. Wouldn't more such detailed explanation in the user guide be helpful? As written, it does nothing helpful for novice users like myself and, in effect, does a disservice to the product. At the least, it could have shortened or eliminated this thread that we have all wasted a lot of time on.

Anyway, this new-to-me setup will be a real help from now on and I really appreciate your taking the effort and time to understand my problem and set me on the right path. I am sure this will make PDF-XCE an even more productive tool more me and I thank you for that.

[PHK]

From the discussions above, I am inclined to infer that what is going on when I Split Pages and Merge Pages is that the new pages that get created in Spllit Pages are more than what you see. Each page has an inaccessible content layer that is the same size as the original page but only the rectangle of the selected area is visible and looks like a page that is only the size of the visible rectangle. What you see is not what you get. Depending on what settings the user has elected to do in the Split Pages process,the content outside the rectangle will either be deleted altogether or preserved, but invisible, in this hidden layer.

The default setting is this:

image(2).png

which will produce these hidden layers with all the content of the the original page for each of the new pages. In my Bose manual example above where the Split Pages process created 20 new pages, you will have 20 hidden layers each with all the content, if mostly invisible, of the original page. In other words, 20 times the content of which 19/20ths is invisible. This will produce horrendous file bloat and lays the ground for ghost search results as I have illustrated above. If these pages are merged, you can produce very difficult-to-manage pages and massive content redundancy. The problem can be a huge one if the user is splitting up a large sheet with a lot of content into many smaller pages as each new page will have tons of unwanted content on its invisible content layer.

I have no idea what the logic is in making these settings the default. Nor do I see any provision for changing those default settings in Preferences.

Daniel's recommendations above

image(3).png

solves the problem. But one must remember to check those settings and probably change them each time one does Split Pages.

But why is the fact that hidden content layers are being created (if I am correct about this) explained in the user guide? Wouldn't it be better for users to have access to information about what is going on "under the hood" or "behind the curtain"?

In my personal experience over the past 18 months, I have created some unnecessarily Frankensteinian files and I have wasted many hours along the way only because I did not really know how the Split Pages function was working. I cannot go back and "unscramble the omlet."

[Daniel - PDF-XChange]

Hello, PHK

At the very least, I can confirm that no additional "layers" should be made with this process, what you are seeing is a single page being duplicated, and all of its content doing the same, then each version of that page is cropped to show a unique area. The reason why these are the default settings is because if the content is deleted, it is irrecoverable once you save the file. Should a mistake be made, or a page be split at an inopportune location (such as through the middle of a paragraph) choosing not to preserve the original content can lead to other issues.

As an example to match your use case, let us assume that you have page which is 4x5 sheets. One night you are working late and mistakenly use split pages to cut it up into 5x4 sheets, only realizing after you saved that you made this cut incorrectly. Now, when you merge the pages back together, any text or images that were on an edge, may have been sliced in a way that required us to remove the text formatting, or recompress the image, and delete the remainder. You can end up with lines of text that visually look like normal text, but are instead two rows of half-letter "shapes". While the defaults can certainly lead to file bloat if used as you are doing, they allow the user to actively make the decision if they wish to remove/delete the extra content themselves (also, the intent behind the split pages function is not exactly for the pages to be "re-merged" again and again repeatedly, which complicates this case a tad)

I should also mention that you said you need to set these options every time? The Split pages function should remember the last used settings, even between system/app restarts. If you are needing to manually re-apply these settings, you may want to uninstall, restart the device, and then reinstall the application, to try to fix permissions issues that could be preventing us from saving your settings.

Kind regards,