DBI (version 1.2.2)

dbGetQuery: Retrieve results from a query

Description

Returns the result of a query as a data frame. dbGetQuery() comes with a default implementation (which should work with most backends) that calls dbSendQuery(), then dbFetch(), ensuring that the result is always freed by dbClearResult(). For retrieving chunked/paged results or for passing query parameters, see dbSendQuery(), in particular the "The data retrieval flow" section. For retrieving results as an Arrow object, see dbGetQueryArrow().

DBI:::methods_as_rd("dbGetQuery")

Usage

dbGetQuery(conn, statement, ...)

Value

dbGetQuery() always returns a data.frame, with as many rows as records were fetched and as many columns as fields in the result set, even if the result is a single value or has one or zero rows.

Arguments

conn

A DBIConnection object, as returned by dbConnect().

statement

a character string containing SQL.

...

Other parameters passed on to methods.

Implementation notes

Subclasses should override this method only if they provide some sort of performance optimization.

Failure modes

An error is raised when issuing a query over a closed or invalid connection, if the syntax of the query is invalid, or if the query is not a non-NA string. If the n argument is not an atomic whole number greater or equal to -1 or Inf, an error is raised, but a subsequent call to dbGetQuery() with proper n argument succeeds.

Additional arguments

The following arguments are not part of the dbGetQuery() generic (to improve compatibility across backends) but are part of the DBI specification:

• n (default: -1)

• params (default: NULL)

• immediate (default: NULL)

They must be provided as named arguments. See the "Specification" and "Value" sections for details on their usage.

Specification

A column named row_names is treated like any other column.

The n argument specifies the number of rows to be fetched. If omitted, fetching multi-row queries with one or more columns returns the entire result. A value of Inf for the n argument is supported and also returns the full result. If more rows than available are fetched (by passing a too large value for n), the result is returned in full without warning. If zero rows are requested, the columns of the data frame are still fully typed. Fetching fewer rows than available is permitted, no warning is issued.

The param argument allows passing query parameters, see dbBind() for details.

Specification for the <code>immediate</code> argument

The immediate argument supports distinguishing between "direct" and "prepared" APIs offered by many database drivers. Passing immediate = TRUE leads to immediate execution of the query or statement, via the "direct" API (if supported by the driver). The default NULL means that the backend should choose whatever API makes the most sense for the database, and (if relevant) tries the other API if the first attempt fails. A successful second attempt should result in a message that suggests passing the correct immediate argument. Examples for possible behaviors:

1. DBI backend defaults to immediate = TRUE internally

   1. A query without parameters is passed: query is executed

   2. A query with parameters is passed:

      1. params not given: rejected immediately by the database because of a syntax error in the query, the backend tries immediate = FALSE (and gives a message)

      2. params given: query is executed using immediate = FALSE

2. DBI backend defaults to immediate = FALSE internally

   1. A query without parameters is passed:

      1. simple query: query is executed

      2. "special" query (such as setting a config options): fails, the backend tries immediate = TRUE (and gives a message)

   2. A query with parameters is passed:

      1. params not given: waiting for parameters via dbBind()

      2. params given: query is executed

Details

This method is for SELECT queries only (incl. other SQL statements that return a SELECT-alike result, e.g., execution of a stored procedure or data manipulation queries like INSERT INTO ... RETURNING ...). To execute a stored procedure that does not return a result set, use dbExecute().

Some backends may support data manipulation statements through this method for compatibility reasons. However, callers are strongly advised to use dbExecute() for data manipulation statements.

See Also

For updates: dbSendStatement() and dbExecute().

Other DBIConnection generics: DBIConnection-class, dbAppendTable(), dbAppendTableArrow(), dbCreateTable(), dbCreateTableArrow(), dbDataType(), dbDisconnect(), dbExecute(), dbExistsTable(), dbGetException(), dbGetInfo(), dbGetQueryArrow(), dbIsReadOnly(), dbIsValid(), dbListFields(), dbListObjects(), dbListResults(), dbListTables(), dbQuoteIdentifier(), dbReadTable(), dbReadTableArrow(), dbRemoveTable(), dbSendQuery(), dbSendQueryArrow(), dbSendStatement(), dbUnquoteIdentifier(), dbWriteTable(), dbWriteTableArrow()

Other data retrieval generics: dbBind(), dbClearResult(), dbFetch(), dbFetchArrow(), dbFetchArrowChunk(), dbGetQueryArrow(), dbHasCompleted(), dbSendQuery(), dbSendQueryArrow()

Examples

if (FALSE) { # requireNamespace("RSQLite", quietly = TRUE)
con <- dbConnect(RSQLite::SQLite(), ":memory:")

dbWriteTable(con, "mtcars", mtcars)
dbGetQuery(con, "SELECT * FROM mtcars")
dbGetQuery(con, "SELECT * FROM mtcars", n = 6)

# Pass values using the param argument:
# (This query runs eight times, once for each different
# parameter. The resulting rows are combined into a single
# data frame.)
dbGetQuery(
  con,
  "SELECT COUNT(*) FROM mtcars WHERE cyl = ?",
  params = list(1:8)
)

dbDisconnect(con)
}