duckdb (version 0.2.4)

duckdb_read_csv: Reads a CSV file into DuckDB

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

# NOT RUN {
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)
# }