PDF Search Not Working? Test the File, Then the Program

When PDF search is not working on a file that has text, decide first whether the fault sits in the file or in the program. Copy a paragraph from the page and paste it into a text editor. Readable words mean the file is sound, and the program, its options or its preview window is what needs attention.

That split saves the most time, because the two branches have nothing in common. A file fault is repaired by getting or making a better file. A program fault is repaired by changing a setting, opening the file somewhere else, or fixing an index, and no amount of OCR will touch it.

Copy a caption into a text editor and look for the petitioner, the respondent, and the docket number. Set that paste beside the page image, then search the same words in the reader. If the paste holds the trial court name and the search returns nothing, the program is narrowing the query. An order page that neither pastes nor searches has no text on that sheet.

Is PDF search failing because of the file or the program?

PDF search fails because of the file when the page holds no characters, or holds the wrong ones. It fails because of the program when the characters are there and the program does not reach them or does not match them. Pasting a paragraph into a plain text editor shows you which case you have.

The paste test

Save the PDF to your drive and open it in a full PDF program. On a page where the word you searched for is plainly printed, copy the paragraph that contains it, paste it into a plain text editor such as Notepad or TextEdit, and look for your word in what arrived. The copy and paste check is one of the tests in the guide to confirming a PDF is really searchable; here it is used to place the fault.

Test the page that failed, and not page one. A file assembled from several sources can have typed pages and pictures side by side, and a result from the cover letter says nothing about the exhibit behind it.

The paste test works because Find and Copy draw on the same stored characters. The post on scanning a PDF for text explains the extraction and matching steps under how a PDF reader scans a file for text.

Why does a PDF search in one program and not in another?

A PDF searches in one program and not in another because each program decides for itself which parts of the file to read and how to join stored characters into words. The file is identical. The programs differ in scope, in how they guess at spaces, and sometimes in which copy of the file they have open.

They read different parts of the file

A PDF carries more than page text. It can hold comments, bookmarks, form fields, layers and attached files. Adobe's help page on searching PDFs in Acrobat says that "Acrobat searches the PDF body text, layers, form fields, and digital signatures", and adds that you "can also include bookmarks and comments in the search". Another program may draw the line elsewhere. If the word you are hunting is typed into a form field or sits in a sticky note, a viewer that reads only page text will pass over it while Acrobat finds it. Each program's own documentation is the place to learn its scope.

They assemble words differently

A PDF may store letters by position without recording where the spaces go. The program infers the spaces. Two programs can look at the same letters and one will read "Calloway Trust" while the other reads "CallowayTrust" or "Callo way Trust". A phrase search succeeds in the first and fails in the second. A search for a single distinctive word is the workaround, and the comparison between two programs is the proof that spacing is the cause.

Editions of one product differ

The desktop, web and mobile editions of a product do not always share features. Adobe's instructions for advanced search begin "Open Acrobat on your desktop (not in a web browser)". If a colleague's search works and yours does not, compare which edition each of you has open before comparing anything else.

They may not have the same file open

Before blaming a program, confirm both are showing the same file. One window may hold the searchable copy and the other the scan it was made from, opened from an old email or a download folder. Check the full path in each title bar or in the document properties, and compare the file sizes.

Searching a PDF inside an email preview, a browser tab or a document system

A preview pane, a browser tab and a document system's viewer are each a separate PDF program with its own search, or with none. A failed search there tells you about that viewer. Save the file, open it in a full PDF program, and repeat the search before you conclude anything about the file.

Email previews

A preview shows an attachment without opening it in a PDF program. Whether your mail program's preview searches the attachment's stored text is something its documentation may not state, and none of the vendor pages opened for this post say it one way or the other. So test it: press the search shortcut while the preview is showing and notice which box the cursor lands in. If it is the mailbox search field, you are searching your email, not the attachment.

Browser tabs

Browsers ship with their own PDF viewers. Microsoft's documentation for the PDF reader built into Microsoft Edge lists search among its basic reading and navigation features. A browser's viewer is still a different program from your desktop PDF editor, and it can assemble words differently.

A second browser case looks the same and is not. Some web applications show a PDF by sending your browser a picture of each page. The address bar shows a web page, not a file ending in .pdf. The browser's find box then searches the web page around the pictures and reports nothing, however good the file's text is. Download the file and open it directly.

Document systems and review platforms

A document management system or review platform can have two search functions that are easy to confuse. One is the system-wide search box, which consults an index the system built earlier. The other is the find control inside the system's own viewer. A word missing from the first is an indexing question for your administrator. A word missing from the second is a viewer question, answered by downloading the file and searching it in a full program.

Reader options that silently narrow a PDF search

Options for whole words, matching case and search scope narrow a PDF search without announcing themselves in the results. When a find box returns nothing for a word you can see, read every checkbox and menu beside it before you suspect the file, then clear the box and type the word again.

Adobe's page on searching PDFs describes two options to look at first. Of "Whole words" it says the option "locates exact matches for the word you enter, excluding partial matches". Of "Case sensitive" it says the option "searches for words with matching capitalization", so that "searching Web won't include web or WEB". With either one ticked, a search behaves differently from the one you ran last week, and the results list does not say why.

Things to check beside the find box

• Whole words. With it on, "indemn" finds nothing even in a contract full of indemnities. It also fails wherever the stored text has lost a space, because "theTenant" is one long word to the program.
• Match case. With it on, "calloway" misses "Calloway" and "CALLOWAY".
• Scope. A search set to a folder, a page range or the current selection reports on that, not on the open document.
• The search term itself. A term pasted from an email can carry a trailing space, a curly apostrophe or a line break. Delete the term and type it by hand.
• Leftovers. Check whether your program has kept a term or an option from your last search. Whether options persist between sessions depends on the program, so look every time until you know.
• Results that are there but hidden. Some programs jump to the first hit only after you press Enter, and some open a results panel off to the side. Confirm you are looking where the results appear.

Options interact. Adobe's page notes that in its advanced search, stemming "isn't available if either Whole Words Only or Case-Sensitive is selected", so one tick can switch off another feature you were counting on.

What each option does to a match, and when to use it on purpose, is laid out in the post on scanning a PDF for text under exact match, case, whole words and phrases. For fault-finding, the rule is simpler: turn everything off, search for one short distinctive word, and add options back one at a time.

Why does the computer's file search not find words inside a searchable PDF?

The computer's file search consults an index built in advance and does not open your PDFs when you type. A searchable PDF stays invisible to it when the folder is outside the indexed locations, when PDF files are indexed by properties only, or when the index has not caught up with a new file.

Microsoft's support page on search indexing in Windows says each file type can be indexed at one of two levels: "either properties only, or properties and content. For properties only, indexing will not look at the contents of the file or make the contents searchable." What an index takes from a PDF, and where the Windows settings are, is set out in the post on folder search under what a desktop or document-system index reads from a PDF.

For fault-finding on a Windows computer, that leaves three questions.

1. Is the folder indexed? A matter folder on a network share or a synced drive may not be among the indexed locations.
2. How is the PDF file type set? The Microsoft page does not say how the pdf type is set on a given computer, so read your own.
3. Has the index seen the file yet? The page says Windows "opens recently changed files, looks at the changes, and stores the new information in the index". A file saved a minute ago may not be in it. Search for the file's name: if the name is found and a word inside is not, the content was not indexed or has not been yet.

On a Mac, or with a third-party desktop search tool, the same three questions apply, and the answers are in that system's own documentation.

A firm-managed computer may lock these settings. In that case the fix is a request to whoever runs your systems, with the three answers attached.

Two cautions. First, a file search that finds nothing has not searched image-only PDFs at all, and it reports them the same way it reports files that lack the word. Second, an index holds what the file contained when it was indexed. If you replaced a scan with a searchable copy under the same name, the index may still hold the empty version until it revisits the file.

Making a whole folder answer to search, and keeping it that way as files arrive, is the subject of the post on converting PDFs so their text is searchable from the folder.

A three-program test that isolates a PDF search fault

Run the same one-word search in three places: your usual PDF program with all options cleared, a second PDF program from a different maker, and a plain text editor holding text pasted from the page. The pattern of hits and misses across the three names the faulty part.

Setting up

• Save the PDF to your drive, so every program opens the same file.
• Pick one page and one test word printed on it. Choose a distinctive word of six letters or more with no hyphen, apostrophe or accent.
• Count by eye how many times the word appears on that page.

Running it

1. Program A, your usual reader. Close and reopen it. Open the file, clear every search option, type the word, and note the hits on your test page.
2. Program B, a different maker's reader. A browser's built-in viewer will do if you open the saved file in it directly. Type the same word and note the hits.
3. Program C, a text editor. Back in either reader, select the whole test page, copy, and paste into the editor. Use the editor's own find for the word.

Reading the result

A worked example

A paralegal on the fictional matter Dunmore Freight v. Harlow Properties searches a 90-page lease file for "Quarry" and gets nothing, though page 4 shows "4410 Quarry Road" twice. In her usual reader with options cleared, the search still misses. In the browser's viewer, opened on the saved file, it hits twice on page 4. The pasted text reads "4410 Quarry Road" correctly.

Row three applies: the file is sound and her reader is the problem. She looks again and finds the search scope set to a folder from a job the day before. With the scope back on the current document, the search hits twice.

Had the browser also missed and the paste shown "Quany Road", the answer would have been the opposite: a recognition error in the file, and a note to search "Quany" as well until the file is redone.

Whichever row you land in, write one line in the file note saying what you tested and what you found. If the conclusion is that the file needs a new text layer, Schema Law reads an uploaded PDF, PNG or JPEG into text and exports a searchable PDF among its formats, as the how it works section of the homepage describes. Recognition there is strongest on clean print, so run the three-program test again on the new file before you rely on it.