read_csv()로 CSV 파일 읽기

기본 R 대신 readr를 사용하는 이유

library(readr)

# read_csv() vs read.csv():
# 1. Returns a tibble (prints nicely, faster subsetting)
# 2. Does NOT convert strings to factors by default
# 3. Shows column specification message
# 4. Faster for large files

# Example with a simple inline CSV:
df <- read_csv('name,score,grade
Alice,85,B
Bob,92,A
Carol,78,C')

print(df)
print(class(df))

파일 경로에서 read_csv() 사용하기

library(readr)

# Reading from a file path (illustrative — file not present)
# df <- read_csv('data/sales_2024.csv')

# Reading from a URL also works:
# df <- read_csv('https://example.com/data.csv')

# Using a literal inline string for demonstration:
df <- read_csv('region,q1,q2,q3
East,100,120,115
West,200,195,210
North,80,85,90')

print(df)

col_types — 열 유형 지정하기

library(readr)

df <- read_csv('id,date,amount,active
1,2024-01-15,99.99,TRUE
2,2024-02-20,149.50,FALSE
3,2024-03-10,75.00,TRUE',
# Compact: integer, Date, double, logical
col_types = 'iDdl')

print(df)
cat('\nColumn types:\n')
print(sapply(df, class))

명시적으로 유형을 지정하는 cols()

library(readr)

df <- read_csv('id,name,score,date
1,Alice,85.5,2024-01-15
2,Bob,92.0,2024-02-20
3,Carol,78.3,2024-03-10',
col_types = cols(
  id = col_integer(),
  name = col_character(),
  score = col_double(),
  date = col_date(format = '%Y-%m-%d')
))

print(df)
print(class(df$date))

skip 및 n_max 인수

library(readr)

# File with metadata in first 2 lines
# (simulated with literal string)
df <- read_csv('# Generated 2024-01-15
# Source: internal DB
name,score,grade
Alice,85,B
Bob,92,A
Carol,78,C
Dave,88,B',
skip = 2,     # Skip the 2 comment lines
n_max = 3     # Read only 3 data rows
)

print(df)

na 인수 — 결측값 정의하기

library(readr)

df <- read_csv('name,score,city
Alice,85,NYC
Bob,N/A,NULL
Carol,78,.
Dave,-999,Chicago',
na = c('', 'NA', 'N/A', 'NULL', '.', '-999')
)

print(df)
cat('NAs per column:\n')
print(colSums(is.na(df)))

col_names — 사용자 지정 열 이름

library(readr)

# File without headers
df <- read_csv('Alice,85,NYC
Bob,92,LA
Carol,78,Chicago',
col_names = c('name','score','city')
)

print(df)

# auto-names when no header:
df2 <- read_csv('1,2,3
4,5,6', col_names = FALSE)
print(df2)

locale() — 표준이 아닌 형식 처리하기

library(readr)

# European-style CSV: semicolons as delimiters, comma as decimal
# read_csv2() is shorthand for this locale
df <- read_csv2('name;price;qty
Alice;9,99;100
Bob;24,50;50',
locale = locale(decimal_mark=',')
)

print(df)

# For dates in non-ISO format:
parse_date('15/01/2024', format='%d/%m/%Y')

progress 및 show_col_types

library(readr)

# Suppress progress and type messages for production scripts
df <- read_csv(
  'a,b,c
  1,x,TRUE
  2,y,FALSE',
  show_col_types = FALSE,
  progress = FALSE
)

print(df)

Problems() — 구문 분석 실패 진단하기

library(readr)

# Intentional type mismatch: 'N/A' in numeric column
df <- suppressWarnings(read_csv(
  'id,score
  1,85
  2,N/A
  3,92',
  col_types = cols(score = col_double())
))

print(df)
print(problems(df))

map()으로 여러 파일 읽기

library(readr)
library(purrr)
library(dplyr)

# Simulated: read and bind multiple CSV files
files <- c('q1.csv', 'q2.csv', 'q3.csv')

# In practice:
# all_data <- map_df(files, read_csv, show_col_types=FALSE)

# Demonstration with inline data:
q1 <- read_csv('month,sales\nJan,100\nFeb,120', show_col_types=FALSE)
q2 <- read_csv('month,sales\nApr,130\nMay,145', show_col_types=FALSE)

bind_rows(list(q1=q1, q2=q2), .id='quarter')

빠른 확인

복습: readr로 read_csv() 사용하기

library(readr)

# Production-ready read_csv() call
df <- read_csv(
  'name,score,city,active
  Alice,85,NYC,TRUE
  Bob,N/A,LA,FALSE
  Carol,78,NULL,TRUE',
  col_types = cols(
    name = col_character(),
    score = col_double(),
    city = col_character(),
    active = col_logical()
  ),
  na = c('', 'NA', 'N/A', 'NULL'),
  show_col_types = FALSE
)

print(df)