How to Find a Specific String or Word in Files and Directories in Linux

Finding specific strings or words across multiple files in Linux is a common task for developers and system administrators. This article explores several Linux commands to efficiently search for text patterns across files and directories.

Using grep

The grep command is a powerful regular expression search tool that matches text patterns in files. At its basic level, it searches for a string within specified files ?

grep 'string' directory-path/*.*

Example

grep 'config' hadoop-2.6.5/etc/hadoop/*.*

The output shows all files containing the word "config" ?

hadoop-2.6.5/etc/hadoop/capacity-scheduler.xml: <configuration>
hadoop-2.6.5/etc/hadoop/core-site.xml:<configuration>
hadoop-2.6.5/etc/hadoop/hadoop-policy.xml:<configuration>
hadoop-2.6.5/etc/hadoop/hdfs-site.xml:<configuration>
hadoop-2.6.5/etc/hadoop/httpfs-site.xml:<configuration>
hadoop-2.6.5/etc/hadoop/kms-acls.xml:<configuration>
hadoop-2.6.5/etc/hadoop/kms-site.xml:<configuration>
hadoop-2.6.5/etc/hadoop/mapred-site.xml.template:<configuration>
hadoop-2.6.5/etc/hadoop/ssl-client.xml.example:<configuration>
hadoop-2.6.5/etc/hadoop/ssl-server.xml.example:<configuration>
hadoop-2.6.5/etc/hadoop/yarn-site.xml:<configuration>

Using grep -r (Recursive Search)

The -r flag enables recursive search through all subdirectories within the specified path. This is useful for searching through complex directory structures ?

grep -r 'string' directory-path/

Example

grep -r 'done' hadoop-2.6.5/

The recursive search finds "done" across all subdirectories ?

hadoop-2.6.5/sbin/slaves.sh:done
hadoop-2.6.5/sbin/distribute-exclude.sh:done
hadoop-2.6.5/sbin/yarn-daemon.sh:done
hadoop-2.6.5/sbin/kms.sh:done
hadoop-2.6.5/sbin/httpfs.sh:done
hadoop-2.6.5/sbin/refresh-namenodes.sh: done
hadoop-2.6.5/sbin/refresh-namenodes.sh: echo "Refresh of namenodes done."
hadoop-2.6.5/sbin/mr-jobhistory-daemon.sh: done
hadoop-2.6.5/sbin/hadoop-daemon.sh:done

Using egrep for Multiple Words

The egrep command allows searching for multiple words simultaneously using the pipe (|) operator for OR logic ?

egrep -r 'word1|word2' directory-path/

Example

egrep -r 'config|comma' hadoop-2.6.5/

This searches for files containing either "config" or "comma" ?

hadoop-2.6.5/etc/hadoop/capacity-scheduler.xml:<configuration>
hadoop-2.6.5/etc/hadoop/capacity-scheduler.xml:</configuration>
hadoop-2.6.5/etc/hadoop/core-site.xml:<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
hadoop-2.6.5/etc/hadoop/core-site.xml:<configuration>
hadoop-2.6.5/etc/hadoop/core-site.xml:</configuration>
hadoop-2.6.5/etc/hadoop/hadoop-policy.xml: The ACL is a comma-separated list of user and group names.
hadoop-2.6.5/etc/hadoop/hadoop-policy.xml: The ACL is a comma-separated list of user and group names.

Common grep Options

Conclusion

Use grep for basic string searches, grep -r for recursive directory searches, and egrep with pipe notation for multiple word patterns. These tools provide efficient file content searching across Linux systems.