Lecture 2

Stderr

• Where we expect error messages

• Unbuffered stream (information is printed immediately)

• Make sure error message doesn't get mixed up with output

• To redirect:

$ 2 > file.txt

How many words occurs in first 20 lines of file.txt?

$ head -n 20 file.txt > temp
$ wc -w < temp > output.txt

Pipeline

# Takes output from cmd before and redirects it as input in second command
$ head -n 20 file.txt | wc -w > output.txt

Suppose words.txt, words2.txt, etc. each contains a list of words, one per line. Produce a duplicate-free list of all words that occur.

$ cat words*.txt | 

$ sort

• sorts lines in lexicographical order

$ uniq

• removes adjacent duplicate lines

• -c counts occurrences adjacent lines

Can we use output from a program as an argument to another program?

• files.txt contains a list of files on one line

• We want to use that list of files as arguments to a program

# counts num of chars, words, lines present in file.txt
$ cat file.txt | wc

# wc of 2 files cat, file.txt
$ wc cat file.txt 

Embedded Command

$ (command args...)
# eg.
$ wc $(cat file.txt)

# Double quotes suppress globbing pattern
$ echo "Today is $(date). I am $(whoami)"
>> Today is Sep 10 13:30:21 EDT 2019.. I am lauradang.

# Single quotes suppress all embedded commands (prints echo exactly as echo was typed)
$ echo 'Today is $(date). I am $(whoami)'
>> Today is $(date). I am $(whoami)

# can also do no quotes
$ echo Today is $(date). I am $(whoami). *.txt

How to find words or patterns in a file

$ egrep

• extended global regular express print

# General Form:
$ egrep pattern file (can be a list of files)

# Example:
$ egrep cs246 file.txt
>> Prints all lines containing cs246

$ egrep "cs246|CS246" file.txt
>> Print every line that contains cs246 or CS246

# Find | in a file or another special symbol
>> egrep "\|" in file.txt

More Pattern Rules

• (pattern) groups contents together

• [chars] match one instance of any character

  • eg. [a-z] matches any of a-z

  • don't do [A-z] to get all of alphabet (looks for ASCII between capital letters and lowercase letters)

• [^chars] matches anything except these characters

• ? : previous pattern occurs 0 or 1 times

  • eg. [Cc][Ss] ?246 (don't know if ppl do CS246 or CS 246)

• * : 0 or more occurrences of preceeding pattern

  • (cs)*246 matches: 246, cs246, cscs246, cscscs246, etc.

• + : 1 or more occurrences of preceeding patern

• . : match one instance of any character

• .* : match 0 or more instances of ny char

• ^ : matches beginning of line

  • eg. ^cs246 (must begin with cs246)

• $ : matches end of line

  • eg. cs246$ (must end with cs246)

Example: Find lines of even length

^(..)*$

ls -l

• Shows "long" form of files

$ ls -l
# type/permission owner group   size (last modified date) name
>> -rw-r--r-- 1 name name   0 Sep 10 13:32 hello.txt

# r -> read
# w -> write
x -> execute