Similar to Java's and Scala's Iterator
, the BatchCursor
type can
can be used to iterate over the data in a collection, but it cannot
be used to modify the underlying collection.
Similar to Java's and Scala's Iterator
, the BatchCursor
type can
can be used to iterate over the data in a collection, but it cannot
be used to modify the underlying collection.
Inspired by the standard Iterator
, provides a way to efficiently
apply operations such as map
, filter
, collect
on the underlying
collection without such operations having necessarily lazy behavior.
So in other words, when wrapping a standard Array
, an application of map
will copy the data to a new Array
instance with its elements
modified, immediately and is thus having strict (eager) behavior.
In other cases, when wrapping potentially infinite collections, like
Iterable
or Stream
, that's when lazy behavior happens.
Sample:
def sum(cursor: BatchCursor[Int]): Long = {
var sum = 0L
while (cursor.hasNext()) {
sum += cursor.next()
}
sum
}
This class is provided as an alternative to Scala's Iterator because:
- the list of supported operations is smaller
- implementations specialized for primitives are provided to avoid boxing
- depending on the implementation, the behaviour of operators
can be eager (e.g.
map
,filter
), but only in case the source cursor doesn't need to be consumed (if the cursor is backed by an array, then a new array gets created, etc.) - the
recommendedBatchSize
can signal how many batch can be processed in batches
Used in the Iterant implementation.
- Companion
- object
Value members
Abstract methods
Creates a cursor by transforming values produced by the source cursor with a partial function, dropping those values for which the partial function is not defined.
Creates a cursor by transforming values produced by the source cursor with a partial function, dropping those values for which the partial function is not defined.
NOTE: application of this function can be either strict or lazy (depending on the underlying cursor type), but it does not modify the original collection.
- Value Params
- pf
the partial function which filters and maps the cursor.
- Returns
a new cursor which yields each value
x
produced by this cursor for whichpf
is defined
Creates a new cursor that advances this cursor past the
first n
elements, or the length of the cursor,
whichever is smaller.
Creates a new cursor that advances this cursor past the
first n
elements, or the length of the cursor,
whichever is smaller.
- Value Params
- n
the number of elements to drop
- Returns
a cursor which produces all values of the current cursor, except it omits the first
n
values.
Returns an cursor over all the elements of the source cursor
that satisfy the predicate p
. The order of the elements
is preserved.
Returns an cursor over all the elements of the source cursor
that satisfy the predicate p
. The order of the elements
is preserved.
NOTE: application of this function can be either strict or lazy (depending on the underlying cursor type), but it does not modify the original collection.
- Value Params
- p
the predicate used to test values.
- Returns
a cursor which produces those values of this cursor which satisfy the predicate
p
.
Tests whether this cursor can provide another element.
Tests whether this cursor can provide another element.
This method can be side-effecting, depending on the implementation and thus it can also throw exceptions. This is because in certain cases the only way to know if there is a next element or not involves triggering dangerous side-effects.
This method is idempotent until the call to next
happens, meaning that multiple hasNext
calls can be
made and implementations are advised to memoize the
result.
- Returns
true
if a subsequent call to next will yield an element,false
otherwise.
Creates a new cursor that maps all produced values of this cursor to new values using a transformation function.
Creates a new cursor that maps all produced values of this cursor to new values using a transformation function.
NOTE: application of this function can be either strict or lazy (depending on the underlying cursor type), but it does not modify the original collection.
- Value Params
- f
is the transformation function
- Returns
a new cursor which transforms every value produced by this cursor by applying the function
f
to it.
Produces the next element of this iterator.
Produces the next element of this iterator.
This method is side-effecting, as it mutates the internal state of the cursor and can throw exceptions.
- Returns
the next element of this iterator, if
hasNext
istrue
, undefined behavior otherwise (can throw exceptions).
In case this cursor is going to be processed eagerly, in batches then this value should be the recommended batch size for the source cursor.
In case this cursor is going to be processed eagerly, in batches then this value should be the recommended batch size for the source cursor.
Examples:
- if this cursor is iterating over a standard
collection with a finite size, it can be something
generous like
1024
- if it's iterating over a cheap infinite iterator
(e.g.
Iterator.range
), it could be 128. - if it does any sort of I/O or blocking of threads,
then the recommended value is
1
.
Basically the batch size should be adjusted according to how expensive processing this cursor is. If it's a strict collection of a finite size, then it can probably be processed all at once. But if it's a big database result set that can block threads on reads, then it's probably wise to do it one item at a time.
Creates an cursor returning an interval of the values produced by this cursor.
Creates an cursor returning an interval of the values produced by this cursor.
- Value Params
- from
the index of the first element in this cursor which forms part of the slice.
- until
the index of the first element following the slice.
- Returns
a cursor which advances this cursor past the first
from
elements usingdrop
, and then takesuntil - from
elements, usingtake
Creates a new cursor that will only emit the
first n
values of this cursor.
Creates a new cursor that will only emit the
first n
values of this cursor.
- Value Params
- n
is the number of values to take
- Returns
a cursor producing only of the first
n
values of this cursor, or else the whole sequence, if it produces fewer thann
values.
Concrete methods
Applies a binary operator to a start value and all elements of this cursor, going left to right.
Applies a binary operator to a start value and all elements of this cursor, going left to right.
NOTE: applying this function on the cursor will consume it completely.
- Type Params
- R
is the result type of the binary operator.
- Value Params
- initial
is the start value.
- op
the binary operator to apply
- Returns
the result of inserting
op
between consecutive elements of this cursor, going left to right with the start valueinitial
on the left. Returnsinitial
if the cursor is empty.
Returns true
in case our cursor is empty or false
if there
are more elements to process.
Returns true
in case our cursor is empty or false
if there
are more elements to process.
Alias for !cursor.hasNext()
.
Returns true
in case our cursor has more elements
to process or false
if the cursor is empty.
Returns true
in case our cursor has more elements
to process or false
if the cursor is empty.
Alias for hasNext.
Converts this cursor into an Array
,
consuming it in the process.
Converts this cursor into an Array
,
consuming it in the process.