CHARACTER CLASS

[] is called a character class, it means if a single character is in the character class, it's matched.
There is a metacharacter - inside the character class, it means a range.
[0-9] means from 0 to 9, which is [0123456789]
[A-Z] means from A to Z
[a-z] means from a to z

In grep -e, the rank of these ranges are,
0-9 > A-Z > a-z, so you can use [0-z] to represent all alphanum char, and [A-z] for any alphabetic char.
And of course you can use mixture pattern like [0-5b-kD-Z]

Reverse match

Inside a character class, ^ means to reverse match,
like egrep '[^0-z]$' target_file will find out any line in that ends up not with alphanumeric nums.

Alternation

| means for or in regex
() can be used to wrap expressions in many cases, like in python do.
egrep -n ' (I|me) .* (yes|no) ' file will find any line that has word "I" or "me" in it, and in the same line, after arbitary words, "yes" or "no" appears.

The expressions wrapped by () can be a single complete regex!
You can treat them like regex nested in regex

Ignoring case

Ignoring case isn't a part of regex, but many tools offers this feature, in egrep, using -i can enable the ignoring case feature.

Word Boundaries

\< and \> are metasequences, it was used to match the start position and end position of a word.
A start position of a word means the letter before it isn't any alphanumeric chars or _.
An end positon of a word means the letter after it isn't alphanumeric char or _.

This is an advancing feature only support by some kind of softwares.

Optional Items

[char]? means char can be show here or not, its presence is optional.

Thus we can also say ? means previous pattern is present 0 time or once.

Other Quantifiers: Repetition

+ means pattern occurs once or any more times . means pattern occurs any times

interval quantifier As an advance function, pattern{min, max} means pattern should occurs at least min times and at most max times.

?, +, * works on the character before them

Parenthesis and Backreferences

\1, \2, \3 will match the first, second, third parenthesis expressions.

This egrep -nr '\<([A-z]+) +\1\>' will much two continous identical words. Let's analyze it.([A-z]+)means chars from 1 to any number composed with letters, a preceding\<and a following +ensured we match a single world here. Then\1means the previous pattern wrapped in(), a preceding +and a following>` also ensures this is a single word.
Thus we can match two continuous identical word.

The Great Escape

For special meta-chars, you can use \ to convert it into a normal character.
But be aware, in a character class, it won't work.