Linux - Module 8: Pipes, Redirection and REGEX

Module 8Pipes, Redirection and REGEXExam Objective3.2 Searching and Extracting Data from FilesObjective SummaryPiping and redirectionPartial POSIXCommand Line and RedirectionCommand Line PipesThe pipe character ( | ) can be used between two commands to send the output of the first as input to the second:ls /etc | headThe output of ls /etc is sent to head as input.Command Line PipelinesMultiple commands can be combined to form pipelines. The order in which commands are added to the pipeline can affect the output:I/O RedirectionThree Input/Output (I/O) streams associated with every command:Standard Input (STDIN) is normally provided by the user via the keyboard.Standard Output (STDOUT) is the output produced by the command when operating correctly. STDOUT normally appears in the same window as where command executed.Standard Error (STERR) is the is the output produced by the command when an error has occurred. STDOUT normally appears in the same window as where command executed.I/O Redirection SymbolsSummary of redirection possible with the bash shell: /path/to/file (Redirect STDOUT overwriting file)>> /path/to/file (Redirect STDOUT appending file)2> /path/to/file (Redirect STDERR overwriting file)2>> /path/to/file (Redirect STDERR appending file)&> /path/to/file (Redirect STDERR and STDOUT overwriting file)&>> /path/to/file (Redirect STDERR and STDOUT appending file)The null deviceThe null device is represented by the /dev/null file. (Otherwise known as the “Bit Bucket”)This file is very useful in redirection of input and output. This file serves two purposes: any output redirected to /dev/null is discarded./dev/null can be used for input to provide a stream of null values. STDIN, STDOUT, and STDERR STDIN or 0Standard Input (STDIN) normally is provided by the keyboard but can be redirected with the a.txt command redirects the output to the file a.txt. Finally, the command cat a.txt sends the file contents to STDOUT, so the output is shown.Appending STDOUT redirection Using a single arrow > for STDOUT redirection will clobber, or overwrite, the file specified. Using the double arrow >> for STDOUT redirection will either create a new file or append an existing one:STDERR or 2Standard Error (STDERR) is the output of a command after an error has occurred.It is normally sent to the console/terminal where the command is executed.ls /fake is a command that will cause an error to be output to STDERR because the /fake file does not exist.Redirecting STDERR ls /fake 2> /tmp/err.msg is a command that would cause an error to be sent to STDERR which is then redirected to the /tmp/err.msg file. The cat /tmp/err.msg command sends the contents of the file to STDOUT to display the file:Disposing of STDERR ls /fake 2> /dev/null is a command that would cause STDERR to be redirected to the /dev/null file, in effect disposing of the error message. Notice cat /dev/null displays no visible output.Working with STDERR and STDOUTfind is a command that searches the filesystem. It sends output to STDOUT when it successfully locates a file that matches your criteria. It sends output to STDERR when it fails to access a directory. The find command will be used to demonstrate redirecting both STDOUT and STDERR on the following slides. More detail about the find command appears later in this chapter.STDERR and STDOUT ExampleThe following example demonstrates the find command searching recursively the /etc/pki directory for any files matching "*.crt". Two lines of STDERR and two lines of STDOUT messages appear: Isolating STDERR In the next example, the STDOUT output is redirected to the /dev/null file, so the STDERR output alone is sent to the terminal window:Isolating STDOUT In the next example, the STDERR output is now redirected to /dev/null file, so the STDOUT output alone is sent to the terminal window:Redirecting Multiple Streams Separately In the next example, the STDERR output is sent to the crt.err file and the STDOUT output is sent to the crt.txt file: Redirecting Multiple Streams CombinedIn this example, both STDOUT and STERR are redirected into the same file, crt.all:find CommandSearching with find commandThe filesystem has hundreds of directories with thousands of files making finding files challenging.The find command is a powerful tool to be able to search for files in different ways including:namesizedateownershipSyntax of find commandThe find command has the following syntax:find [start_dir] [search_op] [criteria] [result]If the starting directory (start_dir) is not specified, then the current directory is assumed.The search option (search_op) is how the search will be done. For example, use the -name option to search by name.Syntax of find command (cont'd)The search criteria (criteria) is the data to be used with the search option. So, if the search option was -name, then the search criteria would be the name of the file to find.The result option (result) defaults to -print, which will output the names of the files that are found. Other result options can perform actions on the files that are found.Searching by file nameConsider the following command: find /etc/pki -name "*.crt"Begins searching recursively from the /etc/pki directoryOutput any file names that match "*.crt" (anything that ends in ".crt“).Displaying file detailThe option -ls will create output similar to the ls -l command. (show both)The columns output are: inode, blocks used, permissions, link count, user owner, group owner, size, date/time, and file name.Searching by file sizeThe -size option can be used by find to search by its size.Large units can be specified as K, M, G, etc.Using +1M means more than one megabyte.Using -1M means less than one megabyte.Useful options for find commandOptionExampleMeaning-maxdepth-maxdepth 1Only search specified directory and its immediate subdirectories-group-group payrollFind any files owned by the payroll group-iname-iname hostsCase insensitive search by filename-mmin-mmin -10Find any files modified in the last ten minutes or less-type-type fFind only regular files-user-user bobFind any files owned by the user bobless CommandViewing files with less commandThe less command is a pager command designed to display only one page of data at a time.The more command is another pager command that has less features than the less command.Both commands allow the user to move back and forth with movement commands to view one page at a time.The help screen in lessOnce in the less program, pressing the "h" key will display the help screen:less movement commandsAs seen in the help screen, the less command has many movement commands. The most common commands are:MovementKeyWindow forwardSpacebarWindow backwardbLine forwardEnterExitqHelphless searching commandsType / to search from cursor to end of file.Type ? to search from cursor to beginning of file.Type pattern to search and press Enter.If more than one match found, press n to go to next match or N to go to previous match.head or tailFiltering with headThe head command displays the first ten lines of a file by default.The -n option allows for the number of lines to be displayed to be specified.head with negative linesNormally the head command displays the number of lines specified from the top of the file.Using -n with a negative value, indicates how many lines from the bottom to not show.This example shows all lines from /etc/passwd except the last thirty-two.Filtering with tailThe tail command displays the last ten lines of a file by default.The -n option allows for the number of lines to be displayed to be specified:tail with positive linesIf the -n option specifies the number of lines to be displayed with a plus ( + ) prefix, then the tail command interprets that to mean to display from that line number to the end of the file:Following with tailThe tail command is able to monitor changes a file and print them out as they occur by using -f option.System administrators frequently follow log files in order to troubleshoot system problems.The user must terminate the tail command when following with the –f option by using CTRL-C.sort CommandSorting files or inputThe sort command will rearrange its output lines according to one or more fields you specify for sorting.Fields are separated by whitespace, although with the –t option, you can specify the delimiter.The default sort is in ascending order, but you can use the -r option to reverse the sorting of a field.The default sort is a dictionary sort, but you can use the -n option to make it a numeric sort.Example of sortIn the following example, the /etc/passwd file is sorted using a : character as a delimiter, by the fourth field numerically and then the third field numerically in reverse:File StatisticsFile statistics with wc commandThe wc command outputs up to three statistics for each file it is given as an argument.By default, wc displays the lines, words and bytes contained in each file.If provided more than one file, then it also calculates the totals of all files.To view individual statistics, specify -l for lines, -w for words or -c for bytes. Example of wc commandTo analyze the number of lines, words and bytes in the /etc/passwd and /etc/passwd- files, the following wc command could be executed:Using wc with pipesThe wc command is often used with pipes so that the output of a command can be analyzed.Using wc -l as the final command in the pipe will count how many lines of output was produced.For example, to determine how many files and directories are in the /etc directory, you could execute: ls /etc | wc -lcut CommandFiltering with cut commandIf you want to extract columns of text, then the cut command provides two simple techniques:By delimiter, where whitespace is the default. The -d option can let you specify other delimiters and -f is used to indicate which fields to extract.By character position, using the -c option with the range of the column to extract.Example of cut commandThe /etc/passwd file is delimited by colon with these fields: account:password:UID:GID:GECOS:directory:shellTo extract the first, and fifth through seventh fields: cut –d: -f1,5-7 /etc/passwdgrep CommandFiltering with grep commandThe grep command can be used to filter standard input or the contents of a file for lines matching a specified pattern.If you want to see where a pattern, or perhaps a word, appears within a file, then the grep command is useful for that purpose.Common grep optionsOptionPurpose--colorColor the matches found-vReverse (negate) matches-cCount matches-nNumber matching lines-lList matching files-iMatch case insensitive-wMatch pattern as a wordBasic Regular ExpressionsBasic Regular ExpressionsBasic Regular Expressions (BRE) are able to be used with the grep command without requiring an option to use them (unlike Extended Regular Expression show later).The simplest regular expressions are just alphabetic or numeric characters that match themselves.The backslash \ can be used to escape the meaning of regular expression meta-characters, including the backslash itself.BRE: the . example The . (period) character matches exactly one character.The example below shows the grep command matching the 'a' followed by two characters.The results show it matched 'abc'.BRE: the [ ] example The [ ] (brackets) characters are used to match exactly one character.The characters can be listed or given as a range.If the first character listed is ^ (caret), then it means not the characters bracketed.BRE: the * example The * (asterisk) character will match zero or more of the previous character.Matching "a*" is not very useful because it might match zero a's (matches every line). Matching "abcd*" would be more useful, since you would need an "abc" followed by zero or more d's.BRE: the ^ example The ^ (caret) character, when appearing at the beginning of the pattern, means that pattern must appear at the beginning of the line.The ^ not at the beginning of a pattern matches itself.BRE: the $ example The $ (dollar sign), when appearing at the end of the pattern, means that pattern must appear at the end of the line.The $ not at the beginning of a pattern matches itself.BRE: Combining ^ and $ Combining both the ^ and $ characters allows for two special matches:'^$' is a blank line match. '^pattern$ matches if the whole line contains only the "pattern“.Extended Regular ExpressionsExtended Regular ExpressionsThe use of Extended Regular Expressions (ERE) requires the -E option when using the grep command.Extended Regular Expressions can be combined with Basic Regular Expressions.The following are ERE characters: ?, +, and |ERE: the + example The + (plus) character will match one or more of the previous character.Matching "a+" is useful because it can match one or more a's, ensuring only lines that have at least one “a” are matched. ERE: the ? example The ? (question mark) character will optionally match one of the previous character.The ? character is useful for matching characters that only occasionally appear in a word. The following example illustrates this: ERE: the | example The | (vertical bar) character will act like an “or” operator between two regular expressions.This alternation operator is useful to be able to match multiple patterns: The xargs commandThe xargs command helps complex piped command sets execute more efficientlyIt attempts to build the longest command line possible with as many arguments as possibleIt tries to prevent executing the command each time for every argument