duckdb_read_csv: Reads a CSV file into DuckDB

In duckdb: DBI Package for the DuckDB Database Management System

Description

Directly reads a CSV file into DuckDB, tries to detect and create the correct schema for it. This usually is much faster than reading the data into R and writing it to DuckDB.

Usage

duckdb_read_csv(
  conn,
  name,
  files,
  header = TRUE,
  na.strings = "",
  nrow.check = 500,
  delim = ",",
  quote = "\"",
  col.names = NULL,
  lower.case.names = FALSE,
  sep = delim,
  transaction = TRUE,
  ...
)

Arguments

conn	A DuckDB connection, created by dbConnect().
name	The name for the virtual table that is registered or unregistered
files	One or more CSV file names, should all have the same structure though
header	Whether or not the CSV files have a separate header in the first line
na.strings	Which strings in the CSV files should be considered to be NULL
nrow.check	How many rows should be read from the CSV file to figure out data types
delim	Which field separator should be used
quote	Which quote character is used for columns in the CSV file
col.names	Override the detected or generated column names
lower.case.names	Transform column names to lower case
sep	Alias for delim for compatibility
transaction	Should a transaction be used for the entire operation
...	Passed on to read.csv()

Value

The number of rows in the resulted table, invisibly.

Examples

con <- dbConnect(duckdb())

data <- data.frame(a = 1:3, b = letters[1:3])
path <- tempfile(fileext = ".csv")

write.csv(data, path, row.names = FALSE)

duckdb_read_csv(con, "data", path)
dbReadTable(con, "data")

dbDisconnect(con)

duckdb documentation built on Dec. 15, 2020, 5:35 p.m.