Grep just one line, then remove

Question

I'd like to know a command to extract just the value from line 8 of this file, minus the <string> and </string>, in other words output only 3.2.2

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>BuildVersion</key>
    <string>8</string>
    <key>CFBundleShortVersionString</key>
    <string>3.2.2</string>
    <key>CFBundleVersion</key>
    <string>399.12</string>
    <key>ProjectName</key>
    <string>ServerApp</string>
    <key>SourceVersion</key>
    <string>399012000000000</string>
</dict>
</plist>

Your suggestions are much appreciated! Thanks, Dan

Don't make us count, which line is line 8? – Barmar Dec 14 '14 at 23:37 — Barmar, Dec 14 '14 at 23:37
To parse XML, use a proper XML parser... – Gilles Quénot Dec 14 '14 at 23:48 — Gilles Quénot, Dec 14 '14 at 23:48

score 9 · Answer 1 · edited May 23 '17 at 12:29

9

As stated by Steven Penny and the link RegEx match open tags except XHTML self-contained tags, to parse XML, a proper xml parser is required, one of them is xmllint

$ xmllint --xpath '/plist/dict/string[2]/text()' file.xml

or with xmlstarlet :

$ xmlstarlet sel -t -v '/plist/dict/string[2]/text()' file.xml

or with saxon-lint :

$ saxon-lint --xpath '/plist/dict/string[2]/text()' file.xml

And an even better XPath expression if you want the version number after CFBundleShortVersionString :

'//key[text()="CFBundleShortVersionString"]/following-sibling::string[1]/text()'

edited May 23 '17 at 12:29

Community

1
1

answered Dec 14 '14 at 23:37

Gilles Quénot

173,512
41
224
223

up-voted for the last expression - value after `CFBundleShortVersionString` :-) – anishsane Dec 15 '14 at 05:25

score 2 · Accepted Answer · edited May 23 '17 at 10:27

2

awk 'NR==8,$0=$3' FS='[<>]'

Result

3.2.2

Set Field Separator to < or >
If on line 8, print Field 3

RegEx match open tags except XHTML self-contained tags

edited May 23 '17 at 10:27

Community

1
1

answered Dec 14 '14 at 23:37

Zombo

1
62
391
407

1

Thanks Steven, this works beautifully and is the most concise. I'll mark this as answered as soon as the ten minutes arrives... – Dan Dec 14 '14 at 23:40

score 1 · Answer 3 · answered Dec 15 '14 at 00:01

1

xmllint 'myfile'|sed -n '8 s#.*>\([[:digit:].]\{1,\}\)<.*#\1#p'

if 3.2.2 is somewhere in the file and it is a unique value you can try sed like

xmllint 'myfile'|sed -n 's#.*>\(3.2.2\)<.*#\1#p'

answered Dec 15 '14 at 00:01

repzero

8,254
2
18
40

1

What's the point using xmllint and not `sed -n '8 s#.*>\([[:digit:].]\{1,\}\)<.*#\1#p' file` ? Still sounds weird, but less – Gilles Quénot Dec 15 '14 at 00:10
point taken..Nevertheless..you will have the same results. xmllint is user friendly with sed (unlike cat) that is why I like using it..:) – repzero Dec 15 '14 at 00:13
"unlike cat" for opening xml files..I meant – repzero Dec 15 '14 at 00:15
@xorg sed takes files as arguments – Dec 15 '14 at 13:00

score 1 · Answer 4 · answered Dec 15 '14 at 04:25

1

With sed it can be done as below.

$ sed -rn '8s#<[a-z]+>([0-9.]+)</[a-z]+>#\1#p' file.xml
3.2.2

answered Dec 15 '14 at 04:25

Kannan Mohan

1,810
12
15

Grep just one line, then remove

4 Answers4