DelimitedReader

Mojo struct 🡭

DelimitedReader

@memory_only
struct DelimitedReader[R: Reader, P: LinePolicy = DefaultLinePolicy, MAX: Int = 64]

Generic delimited-file reader over a Reader.

Suitable for uncomplicated delimited files without quoting (no quoting, no escaping), FAI, BED, GFF, and similar bioinformatic formats.

The hot path — next_view() / for view in dr.views() — yields a DelimitedView with zero heap allocations per row: the span lives in the reader’s internal buffer and the offsets are stack-allocated.

Call .to_record() on the view when you need the record to outlive the current iteration step; next_record() / for record in dr.records() do this automatically (one BString alloc per row).

for view in dr (i.e. __iter__) defaults to the zero-alloc view path.

Example — filter without allocating on every row: ```mojo from blazeseq.io import DelimitedReader, FileReader from std.pathlib import Path from blazeseq.io import DelimitedRecord

var reader = FileReader(Path("data.tsv"))
var dr = DelimitedReader[FileReader](reader^, has_header=True)
var results = List[DelimitedRecord[64]]()
while dr.has_more():
    var view = dr.next_view()
    if String(view.get_span(2)) == "homo_sapiens":
        results.append(view^.to_record())  # alloc only on match
```

Parameters

• R (Reader)
• P (LinePolicy)
• MAX (Int)

Fields

• lines (LineIterator[R])
• policy (P)

Implemented traits

AnyType, ImplicitlyDestructible, Movable

Methods

__init__

def __init__(out self, var reader: R, delimiter: UInt8 = ord("\t"), has_header: Bool = False)

Args:

• reader (R)
• delimiter (UInt8)
• has_header (Bool)
• self (Self)

Returns:

Self

Raises:

has_more

def has_more(self) -> Bool

Args:

• self (Self)

Returns:

Bool

header

def header(ref self) -> Optional[DelimitedRecord[MAX]]

Copy of the stored header record, if present.

Args:

• self (Self)

Returns:

Optional

next_view

def next_view(mut self) -> DelimitedView[MutExternalOrigin, MAX]

Return the next row as a zero-alloc DelimitedView.

The view borrows from the reader’s internal line buffer and is invalidated on the next call to any advancing method. Call .to_record() to obtain an owned DelimitedRecord.

Raises EOFError when no more records are available.

Args:

• self (Self)

Returns:

DelimitedView

Raises:

next_record

def next_record(mut self) -> DelimitedRecord[MAX]

Return the next row as an owned DelimitedRecord.

Convenience wrapper around next_view().to_record(). Raises EOFError when no more records are available.

Args:

• self (Self)

Returns:

DelimitedRecord

Raises:

views

def views(ref self) -> _DelimitedViewIter[R, P, MAX, origin_of(self)]

Iterator yielding zero-alloc DelimitedViews.

Args:

• self (Self)

Returns:

_DelimitedViewIter

records

def records(ref self) -> _DelimitedRecordIter[R, P, MAX, origin_of(self)]

Iterator yielding owned DelimitedRecords.

Args:

• self (Self)

Returns:

_DelimitedRecordIter

__iter__

def __iter__(ref self) -> _DelimitedViewIter[R, P, MAX, origin_of(self)]

Default iteration yields zero-alloc views.

Args:

• self (Self)

Returns:

_DelimitedViewIter