---
title: "Using Stringr for strings"
output: 
  html_document:
    toc: yes
    code_folding: show
    theme:
      bg: "#202123" # black
      fg: "#B8BCC2" # grey
      primary: "#EA80FC" # purple
      pre: "#99ff66"
      code: "#99ff66"
      span: "#99ff66"
      a: "#66ffcc"
      base_font:
        google: Ubuntu
      heading_font:
        google: Raleway
      version: 3
---

```{r setup, include=FALSE}
if (requireNamespace("thematic")) 
  thematic::thematic_rmd(font = "auto")
```

## Load stringr 

Text are strings and the R package stringr and/ stringi are libraries to help deal with string data as part of data analysis.

Note:*Text string color is red and is not customizable, I tried to have another color while having a dark mode.*

```{r echo=TRUE, warning=FALSE, message=FALSE, include=TRUE}
library(stringi)
library(stringr)
library(tidyverse)

# stringr had built in data
fruit = stringr::fruit
head(fruit)

head(sentences)
```


## Text strings

Messy text 
```{r}
txt = "YOU CAN HAVE ALL CAPS DATA edit edit edit"
text2 = "// YoU nEvEr kNöw //"
text3 = "2-009.89.0."
text4 = "<p> ...Text strings  </p> "
text5 = "|| * * * <br> \n didyouseeThis?! {stringr} z = 6 *9"
text6 = "http://www.link.com/strings"
text7 = "bill total $400,35 [07/08/01]"
folder = c("results-001-2020.csv", "results-002-2020.csv", "results-003-2020.csv", "results-004-2020.csv", "results-005-2020.csv")
```


## Stringr Functions

Will use fruits data and text strings.

### detect matches

Detection methods return Boolean array or index location

- `str_detect(string, pattern) `
- `str_detect(fruit, pattern = "a")`
- `str_detect(txt, pattern = "<h1>")`
- `str_starts(fruit, pattern = "ap" )`

```{r}
str_starts(text2, pattern = "//")
```
More detection functions:

- `str_which()` finds the indexes of strings that contain match
- `str_locate()` locates the positions of pattern match
- `str_locate_all(fruit, pattern = "an")`
- `str_coount() `count the number of matches in string
- `str_count(fruit, pattern = "an")`

```{r}
str_which(fruit, pattern = "ma")
```

### Subset strings

- `str_sub(string, start= , end= )` extract substrings from character vector
- `str_subset(string, pattern= )` returns only string in pattern match
- `str_extract()`/ `str_extract_all()` returns 1st match in string as a vector / returns all matches
- `str_match()` / `str_match_all()`  return 1st pattern match as a matrix

```{r}
str_sub(text2, start = 4, end = 17)

str_subset(fruit, pattern = "bl")

str_extract_all(txt, "edit")

```


### manage string lengths

- `str_length(string)`  returns length (width) of entire strings
- `str_pad(string, width= , side= c('left','right','both') ')` add string apdding
- `str_trunc(string, width= , side= )` truncate a string
- `str_squish(string)` trim whitespace from each end

```{r}
str_length(txt)

str_pad(fruit[1], width=15, side='left')  # pad on left side of 1st element in fruits


sentences[1]
str_trunc(sentences[1], width = 17, side='right')

str_squish(sentences[1])
```


### Mutate strings

- `str_replace(string, pattern= ,replacement= )` / `str_replace_all()`
- `str_to_lower()` converts strings to lowercase
- `str_to_upper()` converts strings to uppercase
- `str_to_title()` convert strings to title case

```{r}
str_replace(text2, pattern = "nEvEr", replacement = "always")

str_to_lower(txt, locale = 'en')

str_to_upper(txt)

str_to_title(txt)
```


### Join & Split strings

- `str_c(string1, string2, sep= )` join multiple strings into single string
- `str_flatten(string, collapse= )` combines into a single string
- `str_dup(string, times= n)` repeat strings by **n** times
- `str_split_fixed(string, pattern= , n= )` /`str_split()` / `str_split_()`
- `str_glue()` create string from strings and {expressions}
- `str_glue_data()` use a dataframe/ list to create a string from strings and {expressions}

```{r}
str_c(letters[1:5], LETTERS[1:5], sep = "^")

str_c(txt, text4, sep = " :: ")

str_flatten(fruit[1:5], collapse = " % ")

str_dup(text3, times = 3)

str_split_fixed(txt, pattern = "YOU", n= 3)

str_glue("Pi is {pi}")
str_glue(text2, text7)

# starwars data is from dplyr
str_glue_data(head(starwars), "starwars has {head(name)}")
```


### Order Strings 

```{r}
str_order(letters[1:10], decreasing = F, numeric = T)

str_sort(letters[10:1])
```


### Helpers

`str_view_all(fruit, pattern = "berry")` returns a page with highlight matches. 

```{r}
# override existing encoding of letter
str_conv(fruit[1:4], encoding = "ISO-8859-1")

str_wrap(txt, width = 5, indent = 2, exdent = 3)
str_wrap(letters, width = 5, indent = 7)
```











## Regular Expressions (RegEx)

This section is **need to know** :

- regular expressions go in " "
- special characters, run `?"'"` in RStudio terminal for full list
    - `\\` = `\`
    - `\"` = "
    - `\n` = new line
- use `writeLines("\\.")` to see what is viewed

### Match character

To match a specific character, use `"\\{char}"` example `"\\!"`

- {char} = . ! ? ( ) `{` `}` `\n` `\t` 
- `\d` (any digit) 
- `\w` (any word character)
- need 4 backslashes for 1 back slash `\\\\` 

Note that the '.' expression returns everything except for new line.

### RegEx Helpers

- `"[:digit:]"` matches digits
- `"[:letters:]"` matches letters
- `"[:lower:]"` matches lowercase letters
- `"[:upper:]"` matches uppercase letters
- `"[:alnum:]"` matches letters and numbers
- `"[:punct:]"` matches punctuation
- `"[:graph:]"` matches letters, numbers, punctuation
- `"[:space;]"` matches space characters 
- `"[:blank:]"` matches space and tab (not new line)



```{r}
expressions = c("abc ABC 123\t!?\\(){}\n")

str_extract_all(text7, "[:digit:]")

str_extract_all(text2, "[:alnum:]")

str_extract_all(text2, "[:lower:]")

str_extract_all(text6, "[:alpha:]")

```


### Alternates

Custom searches of strings

```{r}
# match A or B
# str_match(fruit, "apple|pear")
str_count(text3, "0|9")

# match one of
str_match_all(text3, "[09]")

# match anything but
str_extract_all(text6, "[^http://]")

# match a string between a range
str_extract_all(text4, "[Text - strings]")


```


### Quantifiers 


```{r}
# match 0 or more  
str_match_all(txt, "edit?")
str_match(folder, "2020?")

# match ZERO or more
str_match(text3, "8*")

# match exactly n times
str_match_all(text6, "w[2]")

# match n or more times
str_match_all(txt, "edit[2,]")

# match between n and m times
str_match_all(text3, "0[1,3]")


```




### Anchors !

This is the specialized part

```{r}
# match at start of string
str_count(folder, "^results")

# match at end of string
str_detect(folder, ".csv$")
```

In the folder vector is a list of csv files, to extract each of them properly need to use special formatting. In the pattern `\\d{n}` d is for digit and there are 3 digits after results and 4 digits after that. This is useful when running a `purrr` function. 

```{r}
folder

str_extract_all(folder, "^results-\\d{3}-\\d{4}.csv$")
```










### Look arounds

```{r}
s ="bacad"

# match followed by a char
str_match(s, "a(?=c)")

# match not followed by a char
str_extract_all(s, "a(?!c)")

# match preceded by
str_extract_all(s, "(?<!b)a")
```