# RegExp 筆記

* [咻咻的筆記小站](https://hengxiuxu.blogspot.com/2017/10/regular-expression.html)
* [regex101](https://regex101.com/r/Dyi5q9/3)

* :question: 請幫我找出桃園縣(H), 台南市(D), 嘉義縣(Q)中為女生的身分證字號

## Lookahead lookbehind
> 站在 X 往前看 Y -> lookbehind
> 站在 X 往後看 Y -> lookahead
* Ref: [RegExp 應用： lookahead , lookbehind](https://darkk6.blogspot.com/2017/03/regexp-lookahead-lookbehind.html)
* :question: 在一個字串中找出連續數字 6~8 個」，這邊要注意的是，連續九個的話是不要的。
* Test case:
* `12345 XD Hi12345678ab666666cd987654321`
* want:
* `12345678`
* `666666`
* don't want:
* `98765432`
* `87654321`
* Answer
* `(?<!\d)\d{6,8}(?!\d)`
* `(?<!\d)\d{6,8}`
* look for ``
* `\d{6,8}(?!\d)`
* [reg101](https://regex101.com/r/V6EhJD/1)
* lookahead (要找的 `X` 在前面)
* positive
* `X(?=Y)`
* e.g. :question: `123A4C5B67` 我要找一個數字，且後方必須是 `A` 或 `B`
* [Ans](https://regex101.com/r/YFyovN/1)
* negative
* `X(?!Y)`
* e.g. :question: `123A4C5B67` 我要找一個數字，且後方不能是 `A` 或 `B`
* [Ans](https://regex101.com/r/J3I5Wr/1)
* lookbehind (要找的 `X` 在後面)
* positive
* `(?<=Y)X`
* `?<`: 前面的文字
* `=Y`: matches `Y`
* e.g. :question: `AOIOOXPAOCKNJS` 要找任何一個字，且前方必須是O或P。或也可以這樣來稱呼：「尋找接再 O 或 P 後面的一個文字」
* [Ans](https://regex101.com/r/7EY1Rr/1)
* e.g. :question: 擷取出 `file_XXXXX` 後面的文字 XXXXX
* file_record_transcript.pdf
* file_07241999.pdf
* oldfile_from_bk201.pdf
* testfile_fake.pdf.tmp
* [ans](https://regex101.com/r/HylxeS/1)
* e.g. 找出檔案在幾點幾分創立

                ubuntu@ubuntu_qemu_kvm:~/xxx$ ls -la | grep -P "(?<=Dec 29).*\s"

                drwxrwxr-x  5 ubuntu ubuntu 4096 Dec 29 09:28 .
                drwxr-xr-x 16 ubuntu ubuntu 4096 Dec 29 09:29 ..
                drwxrwxr-x  2 ubuntu ubuntu 4096 Dec 29 09:27 regexp
                -rw-rw-r--  1 ubuntu ubuntu   59 Dec 29 09:28 t.txt

* ans:
* e.g. 找出 `(f *Form) 的 method name`

                func (f *Form) Required(fields ...string) {
                func (f *Form) MaxLength(field string, d int) {
                func (f *Form) PermittedValues(field string, opts ...string) {
                func (f *Form) Valid() bool {

* Ans: [regex101](https://regex101.com/r/YFyovN/2)
:::info
* MacOS 不支援 lookbehind!
* 可用 brew install grep
* 並用 -P set to Perl
* ref: [Ask Different](https://apple.stackexchange.com/questions/256902/on-osx-grep-can-i-use-look-ahead#:~:text=grep%20in%20macOS%20does%20not,which%20does%20support%20your%20regex.)

:::
* negative
* `(?<!Y)X`
* `?<`: 前面的文字
* `!Y`: doesn't match `Y`
# BackTrack
* :question: What is backTrack?
* When condition not match, the state will go back to prev state.
# Greedy and lazy
* [咻咻](https://hengxiuxu.blogspot.com/2017/10/regular-expression.html)
* 找 開頭為 `a` 結尾是 `t`
* [reg101](https://regex101.com/r/P1pSz4/22)
* [javascript.info tutorial](https://javascript.info/regexp-greedy-and-lazy)
* :question: Try to explain FSM of `a.*t` and `a.*?t`
# 自己做 regexp parser
* [Finite State Machines and Regular Expressions](https://www.gamedev.net/tutorials/_/technical/general-programming/finite-state-machines-and-regular-expressions-r3176/)
* contain simple c++ example of FSM in regexp
* [LiveOverflow - Regular expression as Finite-state machine - Short](https://www.youtube.com/watch?v=hprXxJHQVfQ)
* 對 FSM 的 brief introduction