write.dbi.ffdf {ETLUtils}

Write ffdf data to a database table by using a DBI connection.

Description

Write ffdf data to a database table by using a DBI connection. This can for example be used to store large ffdf datasets from R in Oracle, SQLite, MySQL, PostgreSQL, Hive or other SQL databases.
Mark that for very large datasets, these SQL databases might have tools to speed up by bulk loading. You might also consider that as an alternative to using this procedure.

Usage

write.dbi.ffdf(
  x,
  name,
  dbConnect.args = list(drv = NULL, dbname = NULL, username = "", password = ""),
  RECORDBYTES = sum(.rambytes[vmode(x)]),
  BATCHBYTES = getOption("ffbatchbytes"),
  by = NULL,
  VERBOSE = FALSE,
  ...
)

Arguments

x

the ffdf to write to the database

name

character string with the name of the table to store the data in. Passed on to dbWriteTable.

dbConnect.args

a list of arguments to pass to DBI's dbConnect (like drv, dbname, username, password). See the examples.

RECORDBYTES

optional integer scalar representing the bytes needed to process a single row of the ffdf

BATCHBYTES

integer: bytes allowed for the size of the data.frame storing the result of reading one chunk. See documentation in read.table.ffdf for more details.

by

integer passed on to chunk indicating to write to the database in chunks of this size. Overwrites the behaviour of BATCHBYTES and RECORDBYTES.

VERBOSE

logical: TRUE to verbose timings for each processed chunk (default FALSE).

...

optional parameters passed on to dbWriteTable

Details

Opens up the DBI connection using DBI::dbConnect, writes data to the SQL table using DBI::dbWriteTable by extracting the data in batches from the ffdf and appending them to the table.

Value

invisible()

See Also

dbWriteTable, chunk

Examples

require(ff)

##
## Example query using data in sqlite
##
require(RSQLite)
dbfile <- system.file("smalldb.sqlite3", package="ETLUtils")
drv <- dbDriver("SQLite")
query <- "select * from testdata limit 10000"
x <- read.dbi.ffdf(query = query, dbConnect.args = list(drv = drv, dbname = dbfile), 
first.rows = 100, next.rows = 1000, VERBOSE=TRUE)

## copy db in package folder to temp folder as CRAN does not allow writing in package dirs
dbfile <- tempfile(fileext = ".sqlite3")
file.copy(from = system.file("smalldb.sqlite3", package="ETLUtils"), to = dbfile)
Sys.chmod(dbfile, mode = "777")
write.dbi.ffdf(x = x, name = "helloworld", row.names = FALSE, overwrite = TRUE,
  dbConnect.args = list(drv = drv, dbname = dbfile), 
  by = 1000, VERBOSE=TRUE)
test <- read.dbi.ffdf(query = "select * from helloworld", 
  dbConnect.args = list(drv = drv, dbname = dbfile))

## clean up for CRAN
file.remove(dbfile)
## Not run: 
require(ROracle)
write.dbi.ffdf(x = x, name = "hellooracle", row.names = FALSE, overwrite = TRUE,
  dbConnect.args = list(drv = dbDriver("Oracle"), 
                        user = "YourUser", password = "YourPassword", dbname = "Mydatabase"), 
  VERBOSE=TRUE)

## End(Not run)