AWK • SED • GREP: Together we can change anything!

Example code from my MacAdmins Conference at Penn State 2024 presentation.

Battery cycle count

Using awk, sed, and grep to get the battery cycle couunt:

ioreg -r -c "AppleSmartBattery" | grep -w "CycleCount" | awk '{ print $3 }' | sed '/{.*}/d'
  
116

Using awk to get the battery cycle count:

ioreg -r -c "AppleSmartBattery" | awk -F ' = ' '/"CycleCount" = / { print $2 }'

116

Using sed to get the battery cycle count:

ioreg -r -c "AppleSmartBattery" | sed -e '/"CycleCount" =/!d' -e 's/.* = //'

116

Using grep to get the battery cycle count:

ioreg -r -c "AppleSmartBattery" | grep -e "\"CycleCount\" = " | grep -o '\d*'

116

Even shorter:

ioreg -l | grep -e "\"CycleCount\" = " | grep -o '\d*'

116

Demonstrating ed

# from the ed Wikipedia page
# https://en.wikipedia.org/wiki/Ed_(software)

$ed # start running ed
a
ed is the standard Unix text editor.
This is line number two.
.
2i
  
.
,l

ed is the standard Unix text editor.$
$
This is line number two.$

w text

63

3s/two/three/
,l

ed is the standard Unix text editor.$
$
This is line number three.$

w text

65

q
$

Similarities and differences in function

Using grep to return HTTP response code

#!/bin/zsh

function checkResponseCode()	{
	httpErrorCodes="000 No HTTP code received
200 Request successful
201 Request to create or update object successful
400 Bad request
401 Authentication failed
403 Invalid permissions
404 Object/resource not found
409 Conflict
500 Internal server error"
	
	responseCode=${1: -3}
	code=$( grep "$responseCode" <<< "$httpErrorCodes" )
	
	echo "$code"
}

responseCode=$( /usr/bin/curl \
--silent \
--url "https://www.google.com" \
--write-out "%{http_code}" )

checkResponseCode "$responseCode"

Examples of addressing

#!/bin/zsh

poem="Mary had a litte lamb.
Its fleece was white as snow.
And everywhere that Mary went,
The lamb was sure to go."

grep 'Mary' <<< "$poem"

sed -n '/Mary/p' <<< "$poem"

awk '/Mary/ { print $0 }' <<< "$poem"
_____________________________________

Mary had a litte lamb.
And everywhere that Mary went,

When to use each

The useless use of cat vs. each command's ability to read files on its own

cat ~/Desktop/list.txt | grep "tacos"
grep "tacos" ~/Desktop/list.txt

cat ~/Desktop/list.txt | sed -n "tacos/p"
sed -n "tacos/p" ~/Desktop/list.txt

cat ~/Desktop/list.txt | awk '/tacos/ { print $0 }'
awk '/tacos/ { print $0 }' ~/Desktop/list.txt

The useless use of echo vs. here strings

echo "$variable" | grep "tacos"
grep "tacos" <<< "$variable"

echo "$variable" | sed -n "tacos/p"
sed -n "tacos/p" <<< "$variable"

echo "$variable" | awk '/tacos/ { print $0 }'
awk '/tacos/ { print $0 }' <<< "$variable"

Single quotes, double quotes, and no quotes

grep tacos <<< "$variable"
grep 'too many tacos' <<< "$variable"
grep "$variable" ~/Desktop/list.txt

sed -n 'tacos/p' <<< "$variable"
sed -n "$variable/p" ~/Desktop/list.txt

awk '/tacos/ { print $0 }' <<< "$variable"

Syntax of a command statement

program options address/pattern command input

Grep examples

xml="<mobile_device_model>
		<model_name>Watch7,3</model_name>
		<display_name>Apple Watch Series 8</display_name>
</mobile_device_model>
<mobile_device_model>
		<model_name>Watch7,4</model_name>
		<display_name>Apple Watch Series 9</display_name>
</mobile_device_model>"

grep "model_name" <<< "$xml"

	<model_name>Watch7,3</model_name>
	<model_name>Watch7,4</model_name>

xml="<mobile_device_model>
		<model_name>Watch7,3</model_name>
		<display_name>Apple Watch Series 8</display_name>
</mobile_device_model>
<mobile_device_model>
		<model_name>Watch7,4</model_name>
		<display_name>Apple Watch Series 9</display_name>
</mobile_device_model>"

grep --after-context 1 "model_name" <<< "$xml"

	<model_name>Watch7,3</model_name>
	<display_name>Apple Watch Series 8</display_name>
--
	<model_name>Watch7,4</model_name>
	<display_name>Apple Watch Series 9</display_name>

xml="<mobile_device_model>
		<model_name>Watch7,3</model_name>
		<display_name>Apple Watch Series 8</display_name>
</mobile_device_model>
<mobile_device_model>
		<model_name>Watch7,4</model_name>
		<display_name>Apple Watch Series 9</display_name>
</mobile_device_model>"

grep --after-context 1 --line-number "model_name" <<< "$xml"

2:	<model_name>Watch7,3</model_name>
3-	<display_name>Apple Watch Series 8</display_name>
--
6:	<model_name>Watch7,4</model_name>
7-	<display_name>Apple Watch Series 9</display_name>

Sed examples

Replace "Watch7,3" model identifier with "Apple Watch Series 9"

modelNames="Watch7,3
Watch7,4
Watch7,5"

sed 's/Watch7,3/Apple Watch Series 9/' <<< "$modelNames"


Apple Watch Series 9
Watch7,4
Watch7,5

Replace each every model identifier with "Apple Watch Series 9" using a single command for each.

modelNames="Watch7,3
Watch7,4
Watch7,5"

sed 's/Watch7,3/Apple Watch Series 9/ ; s/Watch7,4/Apple Watch Series 9/ ; s/Watch7,5/Apple Watch Series 9/' <<< "$modelNames"

Apple Watch Series 9
Apple Watch Series 9
Apple Watch Series 9

Replace every model identifier with "Apple Watch Series 9" using a regular express to match multiple model identifiers.

modelNames="Watch7,3
Watch7,4
Watch7,5"

sed 's/Watch7,\d/Apple Watch Series 9/' <<< "$modelNames"


Apple Watch Series 9
Apple Watch Series 9
Apple Watch Series 9

Syntax for substitution in sed. This is the only syntax where the command precedes the address.

's/pattern/replacement/'

Syntax for all other sed commands. This follows the normal syntax where the command follows the address.

'/pattern/one-letter-command'

The address '2,4' refers to the range of lines 2 through 4. The command 'd' means delete.

list="Line 1
Line 2
Line 3
Line 4
Line 5"

sed '2,4 d' <<< "$list"

Line 1
Line 5

Instead of deleting the range of lines, write them to a file named "numbersFile.txt". The on-screen output still displays every line evaluated, but the text file only contains lines 2 through 4.

list="Line 1
Line 2
Line 3
Line 4
Line 5"

sed '2,4 w /Users/Shared/numbersFile.txt' <<< "$list"

Line 1
Line 2
Line 3
Line 4
Line 5

Awk example

Create mailing labels for everyone in California from a mailing list

mailingList="Abigail Adams, 100 A Street, Albany, CA 94706
Bob Bright, 200 B Street, Bakersfield, CA 93301
Charlie Cartwright, 300 C Street, Cambridge, NY 12816
Denise Darling, 400 D Street, Dale, NY 14039
Edith Ebbing, 500 E Street, Eagleville, CA 96110"

fixedMailingList=$( sed 's/, /\t/g' <<< "$mailingList" )

awk -F "\t" '/CA/ {
print $1
print $2
print $3 ", " $4 $5
print ""
}' <<< "$fixedMailingList"