Add Reference A Capture In The Regex as a sed til

README.md

@@ -10,7 +10,7 @@ pairing with smart people at Hashrocket.
 
 For a steady stream of TILs, [sign up for my newsletter](https://tinyletter.com/jbranchaud).
 
-_1105 TILs and counting..._
+_1106 TILs and counting..._
 
 ---
 
@@ -956,6 +956,7 @@ _1105 TILs and counting..._
 - [Grab All The Method Names Defined In A Ruby File](sed/grab-all-the-method-names-defined-in-a-ruby-file.md)
 - [OSX sed Does Regex A Bit Different](sed/osx-sed-does-regex-a-bit-different.md)
 - [Output Only Lines Involved In A Substitution](sed/output-only-lines-involved-in-a-substitution.md)
+- [Reference A Capture In The Regex](sed/reference-a-capture-in-the-regex.md)
 - [Use An Alternative Delimiter In A Substitution](sed/use-an-alternative-delimiter-in-a-substitution.md)
 
 ### Shell

sed/reference-a-capture-in-the-regex.md

@@ -0,0 +1,32 @@
+# Reference A Capture In The Regex
+
+You create a capture group in a `sed` regex by wrapping a pattern in `\(` and
+`\)`. Generally, this capture group is referenced in the substitution
+expression with `\1`.
+
+The capture references (e.g. `\1`) can also be used in the regex as part of
+specifying the match.
+
+For instance, we can do a capture of a single digit followed by a reference to
+that capture. That will match any line that has a pair of matching consecutive
+digits.
+
+```bash
+$ seq 111 | sed -n 's/\([[:digit:]]\)\1/&/p'
+11
+22
+33
+44
+55
+66
+77
+88
+99
+100
+110
+111
+```
+
+This also uses `&` in the subex which represents the entire match. The `-n` and
+`/p` combination suppresses printing of lines to only those that have
+substitutions.