RecordBatch
Methods
- __construct() : mixed
- add() : array<int, CorruptedRecordException|Rows>
- flush() : array<int, CorruptedRecordException|Rows>
- Under a declared schema a record with a value that does not cast is a corrupted record, handed up and left out of the rows; under an inferred schema it is a schema error.
- inferred() : self
- of() : Generator<int, CorruptedRecordException|Rows>
- $records added in order, numbered from 1, then flushed.
- refuse() : array<int, CorruptedRecordException|Rows>
- A record its reader refused before it became values: it takes its place in the batch, so the batch hands its refused records out in source order.
Methods
__construct()
public
__construct(Schema $schema, Backend $backend, int<1, max> $batchSize[, null|string $uri = null ][, null|Schema $known = null ]) : mixed
Parameters
- $schema : Schema
- $backend : Backend
- $batchSize : int<1, max>
- $uri : null|string = null
-
the file the records come from; null for a source that is not a file
- $known : null|Schema = null
-
a keyed reader's (JSON, array, sequence) full schema, $schema's columns and those it leaves out; under an inferred schema a key it lacks was first seen past the sample. A grid reader passes none: its header is checked once per file
add()
public
add(array<string|int, mixed> $record, int $offset[, int|null $line = null ][, null|string $raw = null ][, null|array<string|int, mixed> $fields = null ]) : array<int, CorruptedRecordException|Rows>
Parameters
- $record : array<string|int, mixed>
- $offset : int
- $line : int|null = null
- $raw : null|string = null
-
the record as the source holds it; null when the source holds no text of it
- $fields : null|array<string|int, mixed> = null
-
the cells a grid reader split the record into, before keying them by header
Tags
Return values
array<int, CorruptedRecordException|Rows> —a full batch's refused records, then its rows; [] while the batch is filling
flush()
Under a declared schema a record with a value that does not cast is a corrupted record, handed up and left out of the rows; under an inferred schema it is a schema error.
public
flush() : array<int, CorruptedRecordException|Rows>
Tags
Return values
array<int, CorruptedRecordException|Rows> —the records refused, in the order they were added, then the rows of the rest; [] when none
inferred()
public
static inferred(Schema $schema, Backend $backend, int<1, max> $batchSize, string|null $uri, null|int<1, max> $sampleSize[, Schema|null $known = null ]) : self
Parameters
- $schema : Schema
- $backend : Backend
- $batchSize : int<1, max>
- $uri : string|null
- $sampleSize : null|int<1, max>
-
the rows the schema was inferred from; null for every row
- $known : Schema|null = null
Return values
selfof()
$records added in order, numbered from 1, then flushed.
public
of(iterable<string|int, array<string|int, mixed>> $records) : Generator<int, CorruptedRecordException|Rows>
Parameters
- $records : iterable<string|int, array<string|int, mixed>>
Tags
Return values
Generator<int, CorruptedRecordException|Rows>refuse()
A record its reader refused before it became values: it takes its place in the batch, so the batch hands its refused records out in source order.
public
refuse(CorruptedRecordException $corrupted) : array<int, CorruptedRecordException|Rows>
Parameters
- $corrupted : CorruptedRecordException
Return values
array<int, CorruptedRecordException|Rows> —as add()