A set of output parameters for const. More...

#include <gpudb/protocol/insert_records_from_files.h>

Public Member Functions
	InsertRecordsFromFilesResponse ()
	Constructs an InsertRecordsFromFilesResponse object with default parameter values. More...

Public Attributes
std::string	tableName

std::string	typeId

std::string	typeDefinition

std::string	typeLabel

std::map< std::string, std::vector< std::string > >	typeProperties

int64_t	countInserted

int64_t	countSkipped

int64_t	countUpdated

std::map< std::string, std::string >	info

std::vector< std::string >	files

Detailed Description

A set of output parameters for const.

Reads from one or more files and inserts the data into a new or existing table. The source data can be located either in KiFS; on the cluster, accessible to the database; or remotely, accessible via a pre-defined external data source.

For delimited text files, there are two loading schemes: positional and name-based. The name-based loading scheme is enabled when the file has a header present and text_has_header is set to true. In this scheme, the source file(s) field names must match the target table's column names exactly; however, the source file can have more fields than the target table has columns. If error_handling is set to permissive, the source file can have fewer fields than the target table has columns. If the name-based loading scheme is being used, names matching the file header's names may be provided to columns_to_load instead of numbers, but ranges are not supported.

Note: Due to data being loaded in parallel, there is no insertion order guaranteed. For tables with primary keys, in the case of a primary key collision, this means it is indeterminate which record will be inserted first and remain, while the rest of the colliding key records are discarded.

Returns once all files are processed.

Definition at line 1079 of file insert_records_from_files.h.