1 of 13

CSE 391

Regular Expressions�sed

Slides created by Josh Ervin and Hunter Schafer. �Based off slides made by Marty Stepp, Jessica Miller, Ruth Anderson, Brett Wortzman, and Zorah Fung

2 of 13

ROADMAP

Introduction to the command line
Input/output redirection, pipes
More input/output redirection, tee, xargs
Git: Fundamentals
Git: Branches, merging, and remote repositories
Regular expressions
More regular expressions, sed
Users and permissions
Bash scripting
Industry applications

3 of 13

AGENDA

More regular expressions

Negative matches

sed

4 of 13

Syntax	Functionality
[0-9]	Any digit
[^0-9]	Not any digit
^	Beginning of line
$	End of line

classes.txt

CSE142

CSE143

CSE311

CSE391

CSE416

CSE446

CSE507

CHEM142

MATH324

MATH409

MATH461

PHIL322

HIST210

Suppose we have the file classes.txt on the left. Each line contains the name for a single class at UW. What is the full grep command to print all CSE classes which are not at the 300 level?

5 of 13

SED

sed: a stream editor for filtering and transforming text

Similar to tools such as “search and replace” in most word processors such as Google Docs or Microsoft Word

General format

sed -r ‘s/REGEX/TEXT/’ file.txt

This will replace all matches of REGEX in the given file with TEXT and output it to the console.

6 of 13

SED

sed -r ‘s/Taylor/Hunter/’ names.txt

Notes:

The thing we are “replacing” (i.e. Taylor), is a regular expression
By default, sed outputs to the console and does not modify the original file

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

Console output

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Hunter

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

7 of 13

SED

sed -ri.bak ‘s/Taylor/Hunter/’ names.txt

Notes:

If you want sed to change the contents of your file, use the -i (in-place) flag to do so. You may add an extension to create a backup file.

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Hunter

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

8 of 13

SED

sed -r ‘s/Jo/Hunter/’ names.txt

Notes:

By default, sed will only replace the first occurrence, per line, of the regex

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

Console output

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Hunter, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

9 of 13

SED

sed -r ‘s/Jo/Hunter/g’ names.txt

Notes:

If the g option is included, then we perform a global match and replace all matches on each line

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

Console output

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Hunter, Hunter

Lamar, Kendrick

Xiu, Xiu

Mayer, Hunterhn

Legend, Hunterhn

10 of 13

SED

sed -r ‘s/\<Jo\>/Hunter/g’ names.txt

Notes:

Recall the left-most string is a regular expression and can be used for more complex matchings.
Also recall that \< and \> match the beginning and ending of words.

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

Console output

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Hunter, Hunter

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

11 of 13

emails.txt

hschafer@uw.edu

djdan@cs.washington.edu

zorahlf@cs.washington.edu

lruzzo@uw.edu

Suppose we have the file emails.txt on the left. Each line contains the email address of a CS faculty member. What is the full sed command to output to the console all uw.edu addresses replaced with cs.washington.edu addresses?

Console Output

hschafers@cs.washington.edu

djdan@cs.washington.edu

zorahlf@cs.washington.edu

lruzzo@cs.washington.edu

12 of 13

SED

sed -r ‘s/^(.*),(.*)$/\2 \1/g’ names.txt

Notes:

We can match first/last names on the left and right side and then flip them on the right side.

names.txt

Knowles, Zorah

Z, Jay

Grande, Ariana

B, Cardi

Swift, Taylor

West, Kanye

Jo, Jo

Lamar, Kendrick

Xiu, Xiu

Mayer, John

Legend, John

Console output

Zorah Knowles

Jay Z

Ariana Grande

Cardi B

Taylor Swift

Kanye West

Jo Jo

Kendrick Lamar

Xiu Xiu

John Mayer

John Legen

13 of 13

emails.txt

hschafers@uw.edu

djdan@cs.washington.edu

zorahlf@cs.washington.edu

lruzzo@uw.edu

Suppose we have the file emails.txt on the left. Each line contains the email address of a CS faculty member. What is the full sed command to output just the basename of their email address (i.e. before the @ symbol)?

Console Output

hschafers

djdan

zorahlf

lruzzo