Under Linux, I can use grep to search a bunch of
files for a character string. Is there an equivalent
command for searching pdf files?

Grep may sort of also work with pdf files. You might want to also use the
strings command to get "clean" srings. Note: *some* pdf files are just images
(no actual text). These would be PDFs created by scanning a document (not
using OCR). Also, many typesetting programs (TeX/LaTex, word-processos, etc),
might do some typesetting "magic" (eg ligitures, etc.) that might make things
hard for grep.

xpdf includes a text search button as part of its UI.

At Wed, 3 Apr 2024 12:45:20 -0000 (UTC) db <dieterhansbritz@gmail.com> wrote:

>
> Under Linux, I can use grep to search a bunch of
> files for a character string. Is there an equivalent
> command for searching pdf files?
>
>

--
Robert Heller -- Cell: 413-658-7953 GV: 978-633-5364
Deepwoods Software -- Custom Software Services
http://www.deepsoft.com/ -- Linux Administration Services
heller@deepsoft.com -- Webhosting Services

Robert Heller <heller@deepsoft.com> wrote or quoted:
>might do some typesetting "magic" (eg ligitures, etc.) that might make things

"ligatures"

Text in PDFs is sometimes compressed. So one can either use
programs like "Agent Ransack" to search for text in PDFs or
tools like "pdftotext" to first create a text file for every
PDF file and then grep those text files.

db <dieterhansbritz@gmail.com> wrote:

> Under Linux, I can use grep to search a bunch of
> files for a character string. Is there an equivalent
> command for searching pdf files?

You can use pdfgrep (https://pdfgrep.org/) for that. It is
available as a package in Fedora and Debian as well.

Tim

ram@zedat.fu-berlin.de (Stefan Ram) wrote or quoted:
>Text in PDFs is sometimes compressed. So one can either use
>programs like "Agent Ransack" to search for text in PDFs or
>tools like "pdftotext" to first create a text file for every
>PDF file and then grep those text files.

PS: "Agent Ransack" is Windows software. "pdftotext" is also
available for Linux. Converting all PDFs to text files needs
to be done only once, and then search operations on those
text files are faster than scanning the PDF files for text
on every search!

On 3 Apr 2024 14:29:40 GMT, Stefan Ram wrote:

> ram@zedat.fu-berlin.de (Stefan Ram) wrote or quoted:
>>Text in PDFs is sometimes compressed. So one can either use programs
>>like "Agent Ransack" to search for text in PDFs or tools like
>>"pdftotext" to first create a text file for every PDF file and then grep
>>those text files.
>
> PS: "Agent Ransack" is Windows software. "pdftotext" is also available
> for Linux. Converting all PDFs to text files needs to be done only
> once, and then search operations on those text files are faster than
> scanning the PDF files for text on every search!

I should maybe have elaborated a bit. Sometimes I
remember a certain phrase or word but forget which
pdf it is in. With text files I can do
grep blabla *.txt
and I wanted an equivalent. Using pdftotext would
mean using it for every suspect pdf. Since a lot of
pdf files are searchable, I figured that such a
command might exist.
But if there really is a pdfgrep command, that might
do the job. I will do some googling, thanks.

On Wed, 3 Apr 2024 15:19:24 -0000 (UTC), db wrote:

> On 3 Apr 2024 14:29:40 GMT, Stefan Ram wrote:
>
>> ram@zedat.fu-berlin.de (Stefan Ram) wrote or quoted:
>>>Text in PDFs is sometimes compressed. So one can either use programs
>>>like "Agent Ransack" to search for text in PDFs or tools like
>>>"pdftotext" to first create a text file for every PDF file and then
>>>grep those text files.
>>
>> PS: "Agent Ransack" is Windows software. "pdftotext" is also
>> available for Linux. Converting all PDFs to text files needs to be
>> done only once, and then search operations on those text files are
>> faster than scanning the PDF files for text on every search!
>
> I should maybe have elaborated a bit. Sometimes I remember a certain
> phrase or word but forget which pdf it is in. With text files I can do
> grep blabla *.txt and I wanted an equivalent. Using pdftotext would mean
> using it for every suspect pdf. Since a lot of pdf files are searchable,
> I figured that such a command might exist.
> But if there really is a pdfgrep command, that might do the job. I will
> do some googling, thanks.

I installed pdfgrep in my Kubuntu system, but it is
not happy. Although the man file is there, even help
doesn't work:

> pdfgrep --help
terminate called after throwing an instance of 'std::runtime_error'
what(): locale::facet::_S_create_c_locale name not valid
Aborted (core dumped)

??

On 04/04/2024 10:50, db wrote:
[...]
> I installed pdfgrep in my Kubuntu system, but it is
> not happy. Although the man file is there, even help
> doesn't work:

I just installed pdfgrep_2.1.2-1build1_amd64.deb in my Mint 20.1 and it
seems to work OK. What version is the Kubuntu one?

Peter

On Thu, 4 Apr 2024 16:57:49 +0100, Peter Flynn wrote:

> On 04/04/2024 10:50, db wrote:
> [...]
>> I installed pdfgrep in my Kubuntu system, but it is not happy. Although
>> the man file is there, even help doesn't work:
>
> I just installed pdfgrep_2.1.2-1build1_amd64.deb in my Mint 20.1 and it
> seems to work OK. What version is the Kubuntu one?
>
> Peter

The man file for pdfgrep says V. 2.1.1. My Kubuntu
is 23.04.