从文本文件中提取匹配结果的行

Question

I need to extract the filename from a text file whereas the output on the text file doesn't have fonts.

So as you can see from the output file below I need to print out results where they are no fonts after the first results? So only the last result has fonts in this output

Does this make sense - Would Grep, Sed or Awk be the answer

So need a output from the text file below that shows that no fonts are present in that PDf within the **START and **END

******************START***********************
name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
/home/user1/Documents/temp1.pdf
******************END***********************
******************START***********************
name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
/home/user1/Documents/temp2.pdf
******************END***********************
******************START***********************
name                                 type              emb sub uni object ID
------------------------------------ ----------------- --- --- --- ---------
BAAAAA+TimesNewRomanPS-BoldMT        TrueType          yes yes yes     14  0
CAAAAA+TimesNewRomanPSMT             TrueType          yes yes yes      9  0
/home/user3/Documents/temp file.pdf
******************END***********************

Answer 1

This prints any line containing ".pdf" if the previous line starts with - .

[me@home]$ awk '{if (st && match($0,".pdf")){print $0}; st=match($0,"^-")}' in.txt
/home/user1/Documents/temp1.pdf
/home/user1/Documents/temp2.pdf

It is not a generic solution, but will work with the input data you've given. I can imagine several edge cases where this might fail but it's all down to the specifications of your input file.

---

Update

(Based on the script you've posted in the comments below) If what you're trying to do is simply to identify PDF files that have no embedded fonts, this might work:

MAGNUM="/mnt/network/User\ 1\ PDF\ 06.12.11/"
has_no_fonts() {
    COUNT=$(pdffonts "$1" 2> /dev/null | wc -l)
    exit $(( $COUNT - 4 ))
}
export -f has_no_fonts
find "$MAGNUM" -type f -name "*.pdf" -exec bash -c 'has_no_fonts "{}"' \; -print

Here's a breakdown of the script:

• Detecting embedded font count. Would have been simple if pdffonts returned a specific value if no fonts were embedded but that is not so. We therefore count the number of output lines and deduct 2 (header lines) to determine the number of embedded fonts

   COUNT=$(pdffonts "$1" 2> /dev/null | wc -l) # number of output lines # exactly 2 if no fonts # exactly 0 if there are errors exit $(( $COUNT - 2 )) # exit 0 (success) if and only if PDF has no fonts 

• bash function exported so it can be used in subshell.

   export -f has_no_fonts 

• Locate pdf files and only print out name if PDF valid and has no fonts

   find ..... -exec bash -c 'has_no_fonts "{}"' \\; -print ------- ------- | | -exec cannot run bash functions Will only print so run in a bash subshell filename if prev command exit with 0 

If you prefer a one-line, the whole script can be written as:

find "$MAGNUM" -name "*.pdf" \
    -exec bash -c 'exit $(($(pdffonts "{}" 2> /dev/null |wc -l) - 2))' \; -print

Answer 2

This might work for you:

sed -n '/^\*/,//{H;/\*END\*/{x;s/\n/&/6;t;s|[^/]*\([^\n]*\).*|\1|p}}' in.txt
/home/user1/Documents/temp1.pdf
/home/user1/Documents/temp2.pdf

Explanation:

1. Focus on lines between lines beginning with *
2. Store such lines in the hold space (HS).
3. When we reach the closing delimiter swap to the HS.
4. Check for 6 or more newlines ie entries that must have fonts and if so bailout.
5. Delete all non-essential text and print out.

Or at a pinch:

sed -n '/^\*/,//{H;/\*END\*/{x;s|[^/]*-\n\(/[^\n]*\).*|\1|p}}' in.txt