R Tips

TidyR

CSS Color

Snippets

Common R tips & code snippets: dplyr data manipulation, stringr text operations, CSS color names, code cheatsheets, etc.

Author

Pinn Xu

Published

March 27, 2026

6 min read · 1060 EN words · 0 中文字

Top Tips

Tip	Notes
Relative Paths	Keep data and code in the same folder; use relative paths to avoid errors caused by spaces in paths
Add `na.rm = TRUE` to stats	Required for `sum`, `mean`, `sd`, `IQR`, etc.; otherwise returns NA when missing values exist
Clarify unit of observation	Confirm what each row represents; for mixed types (e.g. post/comment), filter first: `filter(type == "post")`
Name computed columns inside `count()`	e.g. `count(month = month(date))`
Distinguish `filter()` from `select()`	`filter(condition)` filters rows; `select(col)` selects columns (filter operators)
Use `nrow()` / `pull()` to extract values	`nrow()` returns row count; `pull()` extracts a single value from a data frame
Use stringr + Regex for text	`str_length()` counts characters; `str_count(text, "\\S+")` counts words
Reassign after adding columns	`data <- data \|> mutate(…)`
`TRUE` / `FALSE` aliases	Equivalent to `T` / `F`
Use RStudio Snippets	Use code templates to reduce repetitive typing

1 Exploring Data

1.1 Dimensions

nrow(data)  # number of rows
ncol(data)  # number of columns
dim(data)   # dimensions: rows & columns

1.2 Filtering

Filter: Count observations (rows) matching a condition

data |> filter(condition) |> nrow()

# Example: data has a type column with only "comment" and "post" values
# How many posts are there in total?

# Method 1: nrow() directly returns the row count; other dplyr functions require pull() to extract the value
data |>
    filter(type == "post") |>
    nrow()

# Method 2:
data |>
    filter(type == "post") |>
    count(type) |>  # swapping count and filter order gives the same result
    pull(n)

# Method 3:
data |>
    filter(type == "post") |>
    summarize(n = n()) |>
    pull(n)

1.3 Most Frequent Category

data |> count(col, sort=T) |> head(1) |> pull(col)

# Example: data has a date column with Date datatype (yyyy-mm-dd)
# Which year had the most posts?

# Method 1:
data |>
    filter(type == "post") |>
    count(year = year(date), sort = TRUE) |>
    head(1) |> # sort=T was specified above, so head(1) works directly
    pull(year)

# Method 2:
data |>
    filter(type == "post") |>
    count(year = year(date)) |> # count produces a two-column df: year and n
    slice_max(n, n = 1) |> # first n is the column name, second is row count — selects the row with max n
    pull(year)

1.4 Min/Max Value

Min/Max value of a column

# When asking about all rows (e.g. "overall activity", "across the whole dataframe")
# No need for pipe; if there's a condition (e.g. "only posts"), filter first
min(data$col, na.rm = TRUE)
max(data$col, na.rm = TRUE)

1.5 Unique Values

Unique values of a column

n_distinct(data$col)  # unique value count
unique(data$col)      # unique values

2 Cleaning Data

2.1 Rounded Mean or Median

summarize(mean = round(mean(col, na.rm = TRUE), 2)) |> pull(mean)

# Example: mean score of posts in 2019, rounded to 2 decimal places
data |>
    filter(type == "post" & year(date) == 2019) |>
    summarize(mean = round(mean(score, na.rm = TRUE), 2)) |>
    pull(mean)

# Example: median score per year, rounded to 2 decimal places,
# sorted by median score descending, returned as a dataframe
data |>
    filter(type == "post") |>
    group_by(year = year(date)) |>
    summarize(
        median_score = round(median(score, na.rm = TRUE), 2),
        .groups = "drop"  # group + summarize requires .groups = "drop"
    ) |>
    arrange(desc(median_score)) # sort descending; can also write arrange(-median_score)

2.2 Add a Column

data <- data |> mutate(new_col = operation_on_col)

# Add a character count column
data <- data |>
    mutate(text_length = str_length(text))

# Add a word count column
data <- data |>
    mutate(word_count = str_count(text, "\\S+"))

2.3 Value at Row Maximum

data |>
    slice_max(col1, n = 1) |>
    pull(col2)

# Example: who is the author of the post with the highest word_count?
data |>
    # we want posts only, but data contains other types — filter first
    filter(type == "post") |>
    slice_max(word_count, n = 1) |>
    pull(author)

2.4 Text Length

str_length(text)

2.5 Word Count

str_count(text, "\\S+")

3 Visualization

3.1 Bar Chart

# Example: bar chart of monthly post count in 2019

# Step 1: prepare data for plotting
bar_data <- data |>
    filter(type == "post" & year(date) == 2019) |>
    count(month = month(date)) # no need for sort = TRUE

# Step 2: plot
ggplot(bar_data, aes(x = factor(month), y = n)) +
    # month is numeric; using it directly makes the x-axis continuous — use factor(month)
    # Note: months are ordered in time, so no need to sort bars by size;
    # for nominal vars like "region", use x = reorder(area, -n) for descending order
    # if there are many regions or long names, use coord_flip() for a horizontal bar chart

    # Use geom_col (not _bar): y is a pre-computed value from the n column,
    # not asking ggplot to count rows per category on the fly
    geom_col(fill = "steelblue") + # choose any color you like

    labs(
        title = "Monthly Post Count in 2019",
        x = "Month",
        y = "Post Count"
    )

3.2 Histogram

# Plot histogram of post scores
ggplot(data |> filter(type == "post"), aes(x = score)) +
    geom_histogram(fill = "steelblue", bins = 30) + # set number of bins
    labs(
        title = "Post Score Distribution",
        x = "Post Score",
        y = "Frequency"
    )

4 Text Analysis

# Example: show top 10 most frequent words in posts as a dataframe, and save the top word as a variable
data |>
    filter(type == "post") |>
    unnest_tokens(word, text) |>
    anti_join(stop_words) |>
    count(word, sort = TRUE) |>
    head(10)

top1 <- data |>
    filter(type == "post") |>
    unnest_tokens(word, text) |>
    anti_join(stop_words) |>
    count(word, sort = TRUE) |>
    head(1) |>
    pull(word)

5 Statistics

5.1 Compute Statistics

round(sd(data$col, na.rm = TRUE), 2)
IQR(data$col, na.rm = TRUE)

5.2 Filter Before Computing

filter(condition) |> summarize(mean = mean(col, na.rm = TRUE)) |> pull(mean)

5.3 Statistical Test + Interpret Results

t.test(col1 ~ col2, data = data)

5.4 Linear Regression + Coefficients

m <- lm(col1 ~ col2, data = data)
coef(m)[[1]]  # intercept
coef(m)[[2]]  # slope

6 More Tips

6.1 Common dplyr Functions

Function	Purpose	Example
`filter(condition)`	Keep rows matching the condition	`filter(type == "post" & score > 10)`
`select(col)`	Select columns by name	`select(col1)`, `col1:col3`, `-col3`
`count(col)`	Count rows per unique value in this column	`count(year = year(date), sort = TRUE)`
`nrow()`	Count current rows	Takes no arguments; typically used with `filter()`: `filter(score > mean(score, na.rm=T)) \|> nrow()`
`slice_max(col, n=rows)`	Select the row(s) with the maximum value	`slice_max(score, n = 10)`
`pull(col)`	Return the value(s) of this column	Typically narrow down to one row with filter / slice_max / head / summarize, then `pull(col)`, e.g.: `summarize(mean_age=mean(age,na.rm=T))\|> pull(mean_age)`
`mutate(new_col=expr)`	Add/modify columns in place	`mutate(perc = col1/col2*100)`
`group_by(col)`	Group (used with summarize/mutate)	`group_by(age)`
`summarize(...)`	Return a summary	`group_by(age) \|>` `summarize(avg = mean(score, na.rm=T), .groups=“drop”)`
`arrange(col)`	Sort (ascending by default)	`arrange(age, desc(score))`
`glimpse(df)`	Inspect structure (≈ `str()`)	`glimpse(data)`
`rename(new=old)`	Rename columns (does not create new columns)	`rename(new_col = old_col)`

group_by + summarize: must add .groups="drop"; keeps grouping and summary columns, compressing rows to one per group group_by + mutate: keeps all columns with row count unchanged; must call ungroup() afterward mean(condition) → proportion satisfying the condition; sum(condition) → count satisfying the condition

Note the difference between nrow() and n(): both count rows, but nrow() takes a data frame as argument; n() is a dplyr context function only usable inside summarize() / mutate() (e.g. summarize(number_of_observation = n())); use nrow() in a pipe

# Example:
# data has a type column with only "post" and "comment" values;
# another column word_count is the word count of each post/comment
# How many posts have fewer than 10 words?
data |>
    filter(type == "post" & word_count < 10) |>
    nrow() # directly returns a number

data |>
    filter(type == "post" & word_count < 10) |>
    summarize(count = n()) |> # summarize() still returns a dataframe
    pull(count) # pull is needed to get the scalar value

6.2 Logical Operators in filter()

Operator	Description
`!` `&` `\|`	not / and / or
`==` `!=`	vector equality / inequality
`>` `>=` `<` `<=`	greater than / ≥ / less than / ≤
`%in%`	whether the left element exists in the right vector (not element-wise comparison)

%in% checks element membership; == matches by position (recycled when lengths differ)

# Example:
# data has a date column with Date datatype (yyyy-mm-dd);
# another column type has "post" and "comment" values;
# How many posts have year 2019 or 2023?

data |>
    filter(type == "post" & year(date) %in% c(2019, 2023)) |>
    nrow()

# 1. Condition 1: type is "post", use ==
# 2. Condition 2: year is 2019 or 2023,
#    first extract the year from date using year() which gets the year from a full date (yyyy-mm-dd);
#    then two options to match 2019 or 2023:
#      1) (year(date) == 2019 | year(date) == 2023)
#      2) year(date) %in% c(2019, 2023)
# 3. Both conditions must be met, combine with &
#    Note: if you use option 1), wrap it in parentheses before combining with &
# 4. Use nrow() to count rows matching the filter condition

# Extension:
# To count all posts from 2019 through 2023, use:
#   year(date) %in% 2019:2023
# because 2019:2023 expands to c(2019, 2020, 2021, 2022, 2023)

6.3 Helper Functions in select()

Note: select() picks columns; filter() picks rows — they are not interchangeable

Syntax	Effect
`select(col1:col3)`	Select col1 through col3
`select(-col3)`	Exclude col3
`select(contains(“x”))`	Column name contains “x”
`select(starts_with(“c”))`	Column name starts with “c”
`select(ends_with(“3”))`	Column name ends with “3”
`select(everything())`	Select all columns

6.4 Regex

# Add a character count column
data <- data |>
    mutate(text_length = str_length(text))

# Add a word count column
data <- data |>
    mutate(word_count = str_count(text, "\\S+"))

Pattern	Meaning	Example	Matches
`.`	any single character	`"a.c"`	“abc”, “a1c”, “a c”
`^`	start of string	`"^The"`	“The dog” but not “See The dog”
`$`	end of string	`"end$"`	“the end” but not “endless”
`*`	preceding character zero or more times	`"ab*c"`	“ac”, “abc”, “abbc”
`+`	preceding character one or more times	`"ab+c"`	“abc”, “abbc” but not “ac”
`?`	preceding character zero or one times	`"colou?r"`	“color”, “colour”
`\\`	escape special character	`"\\."`	literal period “.”

Pattern	Meaning	Example
`[abc]`	matches a, b, or c	`"[aeiou]"` matches vowels
`[a-z]`	any lowercase letter
`[A-Z]`	any uppercase letter
`[0-9]`	any digit	equivalent to `\\d`
`[^abc]`	any character except a, b, c	`[^0-9]` matches non-digits

Shorthand	Meaning	Equivalent
`\\d`	any digit	`[0-9]`
`\\D`	any non-digit	`[^0-9]`
`\\w`	any word character	`[a-zA-Z0-9_]`
`\\W`	any non-word character	`[^a-zA-Z0-9_]`
`\\s`	any whitespace character	space, tab, newline
`\\S`	any non-whitespace character

6.5 stringr Functions

Function	Purpose	Example
`str_detect()`	whether a match exists (TRUE/FALSE)	`str_detect(x, "pattern")`
`str_subset()`	return matching elements	`str_subset(x, "pattern")`
`str_extract()`	extract first match	`str_extract(x, "pattern")`
`str_extract_all()`	extract all matches	`str_extract_all(x, "pattern")`
`str_replace()`	replace first match	`str_replace(x, "pattern", "replacement")`
`str_replace_all()`	replace all matches	`str_replace_all(x, "pattern", "replacement")`
`str_match()`	extract capture groups from first match	`str_match(x, "pattern")`
`str_count()`	count number of matches	`str_count(x, "pattern")`
`str_split()`	split string by pattern	`str_split(x, "pattern")`

6.6 Common Color Names

Color Name	Preview	Color Name	Preview
pink	████	beige	████
lightpink	████	darkseagreen	████
lightcoral	████	mediumseagreen	████
palevioletred	████	forestgreen	████
rosybrown	████	green	████
indianred	████	seagreen	████
firebrick	████	teal	████
tomato	████	lightseagreen	████
coral	████	mediumaquamarine	████
salmon	████	powderblue	████
lightsalmon	████	lightblue	████
sandybrown	████	skyblue	████
orange	████	lightskyblue	████
peru	████	lightsteelblue	████
tan	████	steelblue	████
wheat	████	mediumpurple	████
bisque	████	plum	████
linen	████	thistle	████

6.7 Relative Paths

Relative path. Place your Rmd file and CSV in the same folder; read with read_csv("./exam-data.csv") — ./ means the current folder. Relative paths are recommended to avoid errors caused by spaces in absolute paths.

Absolute path. In Finder, select the file and press ⌥⌘C to copy the absolute path, e.g. data <- read_csv("/Users/username/EM747/mid-exam/exam-data.csv").

6.8 RStudio Snippet

In RStudio > Tools > Edit Code Snippets

R > scroll to the bottom and add new code > Save

For example, a ggplot template:

snippet ggplot
    ggplot(${1:data}, aes(x = ${2:x}, y = ${3:y})) +
        geom_${4:bar}(${5:fill} = "${6:steelblue}") +
        labs(
            title = "${7:title}",
            x = "${8:xlab}",
            y = "${9:ylab}"
        )

After saving, type ggplot in an R code block and select the snippet:

Press Enter and the template is inserted automatically:

↑ Back to Top

--- title: 'R Tips' author: 'Pinn Xu' date: '2026-03-27' # categories: ['R', 'dplyr', 'stringr', 'Regex', 'CSS Color', 'Text Analysis', 'Visualization', 'Snippets'] categories: ['R', 'TidyR', 'CSS Color', 'Snippets'] description: "Common R tips & code snippets: dplyr data manipulation, stringr text operations, CSS color names, code cheatsheets, etc." order: 5 eval: FALSE aliases: - ../r-tips/ --- # Top Tips {.unnumbered} | Tip | Notes | |------|------| | [Relative Paths](#relative-paths) | Keep data and code in the same folder; use relative paths to avoid errors caused by spaces in paths | | Add `na.rm = TRUE` to stats | Required for `sum`, `mean`, `sd`, `IQR`, etc.; otherwise returns NA when missing values exist | | Clarify unit of observation | Confirm what each row represents; for mixed types (e.g. post/comment), filter first: `filter(type == "post")` | | Name computed columns inside `count()` | e.g. <code>count(month = month(date))</code> | | Distinguish `filter()` from `select()` | `filter(condition)` filters rows; `select(col)` selects columns [(filter operators)](#logical-operators-in-filter) | | Use `nrow()` / `pull()` to extract values | `nrow()` returns row count; `pull()` extracts a single value from a data frame | | Use stringr + Regex for text | `str_length()` counts characters; `str_count(text, "\\S+")` counts words | | Reassign after adding columns | <code>data <- data \|> mutate(...)</code> | | `TRUE` / `FALSE` aliases | Equivalent to `T` / `F` | | Use [RStudio Snippets](#rstudio-snippet) | Use code templates to reduce repetitive typing | # Exploring Data ## Dimensions ```{r eval=FALSE} nrow(data) # number of rows ncol(data) # number of columns dim(data) # dimensions: rows & columns ``` ## Filtering > Filter: Count observations (rows) matching a condition ```{r eval=FALSE} data |> filter(condition) |> nrow() ``` ```{r eval=FALSE} # Example: data has a type column with only "comment" and "post" values # How many posts are there in total? # Method 1: nrow() directly returns the row count; other dplyr functions require pull() to extract the value data |> filter(type == "post") |> nrow() # Method 2: data |> filter(type == "post") |> count(type) |> # swapping count and filter order gives the same result pull(n) # Method 3: data |> filter(type == "post") |> summarize(n = n()) |> pull(n) ``` ## Most Frequent Category ```{r eval=FALSE} data |> count(col, sort=T) |> head(1) |> pull(col) ``` ```{r eval=FALSE} # Example: data has a date column with Date datatype (yyyy-mm-dd) # Which year had the most posts? # Method 1: data |> filter(type == "post") |> count(year = year(date), sort = TRUE) |> head(1) |> # sort=T was specified above, so head(1) works directly pull(year) # Method 2: data |> filter(type == "post") |> count(year = year(date)) |> # count produces a two-column df: year and n slice_max(n, n = 1) |> # first n is the column name, second is row count — selects the row with max n pull(year) ``` ## Min/Max Value > Min/Max value of a column ```{r eval=FALSE} # When asking about all rows (e.g. "overall activity", "across the whole dataframe") # No need for pipe; if there's a condition (e.g. "only posts"), filter first min(data$col, na.rm = TRUE) max(data$col, na.rm = TRUE) ``` ## Unique Values > Unique values of a column ```{r eval=FALSE} n_distinct(data$col) # unique value count unique(data$col) # unique values ``` # Cleaning Data ## Rounded Mean or Median ```{r eval=FALSE} summarize(mean = round(mean(col, na.rm = TRUE), 2)) |> pull(mean) ``` ```{r eval=FALSE} # Example: mean score of posts in 2019, rounded to 2 decimal places data |> filter(type == "post" & year(date) == 2019) |> summarize(mean = round(mean(score, na.rm = TRUE), 2)) |> pull(mean) ``` ```{r eval=FALSE} # Example: median score per year, rounded to 2 decimal places, # sorted by median score descending, returned as a dataframe data |> filter(type == "post") |> group_by(year = year(date)) |> summarize( median_score = round(median(score, na.rm = TRUE), 2), .groups = "drop" # group + summarize requires .groups = "drop" ) |> arrange(desc(median_score)) # sort descending; can also write arrange(-median_score) ``` ## Add a Column ```{r eval=FALSE} data <- data |> mutate(new_col = operation_on_col) ``` ```{r eval=FALSE} # Add a character count column data <- data |> mutate(text_length = str_length(text)) # Add a word count column data <- data |> mutate(word_count = str_count(text, "\\S+")) ``` ## Value at Row Maximum ```{r eval=FALSE} data |> slice_max(col1, n = 1) |> pull(col2) ``` ```{r eval=FALSE} # Example: who is the author of the post with the highest word_count? data |> # we want posts only, but data contains other types — filter first filter(type == "post") |> slice_max(word_count, n = 1) |> pull(author) ``` ## Text Length ```{r eval=FALSE} str_length(text) ``` ## Word Count ```{r eval=FALSE} str_count(text, "\\S+") ``` # Visualization ## Bar Chart ```{r eval=FALSE} # Example: bar chart of monthly post count in 2019 # Step 1: prepare data for plotting bar_data <- data |> filter(type == "post" & year(date) == 2019) |> count(month = month(date)) # no need for sort = TRUE # Step 2: plot ggplot(bar_data, aes(x = factor(month), y = n)) + # month is numeric; using it directly makes the x-axis continuous — use factor(month) # Note: months are ordered in time, so no need to sort bars by size; # for nominal vars like "region", use x = reorder(area, -n) for descending order # if there are many regions or long names, use coord_flip() for a horizontal bar chart # Use geom_col (not _bar): y is a pre-computed value from the n column, # not asking ggplot to count rows per category on the fly geom_col(fill = "steelblue") + # choose any color you like labs( title = "Monthly Post Count in 2019", x = "Month", y = "Post Count" ) ``` ## Histogram ```{r eval=FALSE} # Plot histogram of post scores ggplot(data |> filter(type == "post"), aes(x = score)) + geom_histogram(fill = "steelblue", bins = 30) + # set number of bins labs( title = "Post Score Distribution", x = "Post Score", y = "Frequency" ) ``` # Text Analysis ```{r eval=FALSE} # Example: show top 10 most frequent words in posts as a dataframe, and save the top word as a variable data |> filter(type == "post") |> unnest_tokens(word, text) |> anti_join(stop_words) |> count(word, sort = TRUE) |> head(10) top1 <- data |> filter(type == "post") |> unnest_tokens(word, text) |> anti_join(stop_words) |> count(word, sort = TRUE) |> head(1) |> pull(word) ``` # Statistics ## Compute Statistics ```{r eval=FALSE} round(sd(data$col, na.rm = TRUE), 2) IQR(data$col, na.rm = TRUE) ``` ## Filter Before Computing ```{r eval=FALSE} filter(condition) |> summarize(mean = mean(col, na.rm = TRUE)) |> pull(mean) ``` ## Statistical Test + Interpret Results ```{r eval=FALSE} t.test(col1 ~ col2, data = data) ``` ## Linear Regression + Coefficients ```{r eval=FALSE} m <- lm(col1 ~ col2, data = data) coef(m)[[1]] # intercept coef(m)[[2]] # slope ``` # More Tips ## Common dplyr Functions | Function | Purpose | Example | |------|------|------| | `filter(condition)` | Keep rows matching the condition | `filter(type == "post" & score > 10)` | | `select(col)` | Select columns by name | `select(col1)`, `col1:col3`, `-col3` | | `count(col)` | Count rows per unique value in this column | <code>count(year = year(date), sort = TRUE)</code> | | `nrow()` | Count current rows | Takes no arguments; typically used with `filter()`: <code>filter(score > mean(score, na.rm=T)) \|> nrow()</code> | | `slice_max(col, n=rows)` | Select the row(s) with the maximum value | <code>slice_max(score, n = 10)</code> | | `pull(col)` | Return the value(s) of this column | Typically narrow down to one row with filter / slice_max / head / summarize, then `pull(col)`, e.g.: <code>summarize(mean_age=mean(age,na.rm=T))\|> pull(mean_age)</code> | | `mutate(new_col=expr)` | Add/modify columns in place | `mutate(perc = col1/col2*100)` | | `group_by(col)` | Group (used with summarize/mutate) | `group_by(age)` | | `summarize(...)` | Return a summary | <code>group_by(age) \|></code> <code>summarize(avg = mean(score, na.rm=T), .groups="drop")</code> | | `arrange(col)` | Sort (ascending by default) | <code>arrange(age, desc(score))</code> | | `glimpse(df)` | Inspect structure (≈ `str()`) | `glimpse(data)` | | `rename(new=old)` | Rename columns (does not create new columns) | `rename(new_col = old_col)` | > `group_by` + `summarize`: must add `.groups="drop"`; keeps grouping and summary columns, compressing rows to one per group > `group_by` + `mutate`: keeps all columns with row count unchanged; must call `ungroup()` afterward > `mean(condition)` → **proportion** satisfying the condition; `sum(condition)` → **count** satisfying the condition > Note the difference between `nrow()` and `n()`: both count rows, but `nrow()` takes a data frame as argument; `n()` is a dplyr context function only usable inside `summarize()` / `mutate()` (e.g. `summarize(number_of_observation = n())`); use `nrow()` in a pipe ```{r eval=FALSE} # Example: # data has a type column with only "post" and "comment" values; # another column word_count is the word count of each post/comment # How many posts have fewer than 10 words? data |> filter(type == "post" & word_count < 10) |> nrow() # directly returns a number data |> filter(type == "post" & word_count < 10) |> summarize(count = n()) |> # summarize() still returns a dataframe pull(count) # pull is needed to get the scalar value ``` ## Logical Operators in filter() | Operator | Description | |-----|-----------------| | `!` `&` <code>\|</code> | not / and / or | | `==` `!=` | vector equality / inequality | | `>` `>=` `<` `<=` | greater than / ≥ / less than / ≤ | | `%in%` | whether the left element exists in the right vector (not element-wise comparison) | > `%in%` checks element membership; `==` matches by position (recycled when lengths differ) ```{r eval=FALSE} # Example: # data has a date column with Date datatype (yyyy-mm-dd); # another column type has "post" and "comment" values; # How many posts have year 2019 or 2023? data |> filter(type == "post" & year(date) %in% c(2019, 2023)) |> nrow() # 1. Condition 1: type is "post", use == # 2. Condition 2: year is 2019 or 2023, # first extract the year from date using year() which gets the year from a full date (yyyy-mm-dd); # then two options to match 2019 or 2023: # 1) (year(date) == 2019 | year(date) == 2023) # 2) year(date) %in% c(2019, 2023) # 3. Both conditions must be met, combine with & # Note: if you use option 1), wrap it in parentheses before combining with & # 4. Use nrow() to count rows matching the filter condition # Extension: # To count all posts from 2019 through 2023, use: # year(date) %in% 2019:2023 # because 2019:2023 expands to c(2019, 2020, 2021, 2022, 2023) ``` ## Helper Functions in select() Note: `select()` picks **columns**; `filter()` picks **rows** — they are not interchangeable | Syntax | Effect | |------|------| | <code>select(col1:col3)</code> | Select col1 through col3 | | <code>select(-col3)</code> | Exclude col3 | | <code>select(contains("x"))</code> | Column name contains "x" | | <code>select(starts_with("c"))</code> | Column name starts with "c" | | <code>select(ends_with("3"))</code> | Column name ends with "3" | | <code>select(everything())</code> | Select all columns | ## Regex ```{r eval=FALSE} # Add a character count column data <- data |> mutate(text_length = str_length(text)) # Add a word count column data <- data |> mutate(word_count = str_count(text, "\\S+")) ``` | Pattern | Meaning | Example | Matches | |:---:|-------|:---:|------------| | `.` | any single character | `"a.c"` | "abc", "a1c", "a c" | | `^` | start of string | `"^The"` | "The dog" but not "See The dog" | | `$` | end of string | `"end$"` | "the end" but not "endless" | | `*` | preceding character zero or more times | `"ab*c"` | "ac", "abc", "abbc" | | `+` | preceding character one or more times | `"ab+c"` | "abc", "abbc" but not "ac" | | `?` | preceding character zero or one times | `"colou?r"` | "color", "colour" | | `\\` | escape special character | `"\\."` | literal period "." | | Pattern | Meaning | Example | |:--------:|:-------|:--------| | `[abc]` | matches a, b, or c | `"[aeiou]"` matches vowels | | `[a-z]` | any lowercase letter | | | `[A-Z]` | any uppercase letter | | | `[0-9]` | any digit | equivalent to `\\d` | | `[^abc]` | any character except a, b, c | `[^0-9]` matches non-digits | : {tbl-colwidths=[30,30,40]} | Shorthand | Meaning | Equivalent | |:---:|---------|------------| | `\\d` | any digit | `[0-9]` | | `\\D` | any non-digit | `[^0-9]` | | `\\w` | any word character | `[a-zA-Z0-9_]` | | `\\W` | any non-word character | `[^a-zA-Z0-9_]` | | `\\s` | any whitespace character | space, tab, newline | | `\\S` | any non-whitespace character | | : {tbl-colwidths=[30,30,40]} ## stringr Functions | Function | Purpose | Example | |------|:----------:|------------| | `str_detect()` | whether a match exists (TRUE/FALSE) | `str_detect(x, "pattern")` | | `str_subset()` | return matching elements | `str_subset(x, "pattern")` | | `str_extract()` | extract first match | `str_extract(x, "pattern")` | | `str_extract_all()` | extract all matches | `str_extract_all(x, "pattern")` | | `str_replace()` | replace first match | `str_replace(x, "pattern", "replacement")` | | `str_replace_all()` | replace all matches | `str_replace_all(x, "pattern", "replacement")` | | `str_match()` | extract capture groups from first match | `str_match(x, "pattern")` | | `str_count()` | count number of matches | `str_count(x, "pattern")` | | `str_split()` | split string by pattern | `str_split(x, "pattern")` | ## Common Color Names | Color Name | Preview | Color Name | Preview | |---------|:----:|---------|:----:| | pink | ████ | beige | ████ | | lightpink | ████ | darkseagreen | ████ | | lightcoral | ████ | mediumseagreen | ████ | | palevioletred | ████ | forestgreen | ████ | | rosybrown | ████ | green | ████ | | indianred | ████ | seagreen | ████ | | firebrick | ████ | teal | ████ | | tomato | ████ | lightseagreen | ████ | | coral | ████ | mediumaquamarine | ████ | | salmon | ████ | powderblue | ████ | | lightsalmon | ████ | lightblue | ████ | | sandybrown | ████ | skyblue | ████ | | orange | ████ | lightskyblue | ████ | | peru | ████ | lightsteelblue | ████ | | tan | ████ | steelblue | ████ | | wheat | ████ | mediumpurple | ████ | | bisque | ████ | plum | ████ | | linen | ████ | thistle | ████ | : {tbl-colwidths=[15,35,15,35]} ## Relative Paths **Relative path.** Place your Rmd file and CSV in the same folder; read with `read_csv("./exam-data.csv")` — `./` means the current folder. Relative paths are recommended to avoid errors caused by spaces in absolute paths. **Absolute path.** In Finder, select the file and press `⌥⌘C` to copy the absolute path, e.g. `data <- read_csv("/Users/username/EM747/mid-exam/exam-data.csv")`. ## RStudio Snippet In RStudio > Tools > Edit Code Snippets ![](../../../src/img/ttl/r-tips/rstudio-snippet-1.png){width="400px" fig-alt="Edit code snippets"} R > scroll to the bottom and add new code > Save ![](../../../src/img/ttl/r-tips/rstudio-snippet-2.png){fig-alt="Procedure of creating snippets in RStudio"} For example, a ggplot template: ```{r eval=FALSE} snippet ggplot ggplot(${1:data}, aes(x = ${2:x}, y = ${3:y})) + geom_${4:bar}(${5:fill} = "${6:steelblue}") + labs( title = "${7:title}", x = "${8:xlab}", y = "${9:ylab}" ) ``` After saving, type `ggplot` in an R code block and select the snippet: ![](../../../src/img/ttl/r-tips/rstudio-snippet-3.png){width="550px" fig-alt="type snippet alias in a code cell"} Press Enter and the template is inserted automatically: ![](../../../src/img/ttl/r-tips/rstudio-snippet-4.png){width="350px" fig-alt="inserted snippets"} [↑ Back to Top](#)