The Linux shell writes regex pattern recognition string words __linux

string: AA bb cc DD

Requirements: The word in the output string

Pattern 1:

*@*:~$ echo AA bb cc DD |GREP//b//w*//b

RESULT:AA BB cc DD

Pattern 2: Substituting the + number for *

*@*:~$ echo AA bb cc DD |GREP//b//w+//b

Result: No output

What's the problem, right? + to escape, add a/

Pattern 3:

*@*:~$ echo AA bb cc DD |GREP//b//w/+//b

Result: There is still no output

What's going on here, get it, get it, just know what's going on, look underneath it.

Pattern 4:

*@*:~$ echo AA bb cc DD |GREP//b//w//+//b

RESULT:AA BB cc DD

Finally give a shell command that calculates the number of words

echo AA bb cc dd |SED-E s///b//w//+//b/x/g | SED-E s///s//g | Wc-m (doesn't seem to be right, the output is 5, but actually there are only 4)

I hope you will not doubt