Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support bloom filters in lists/nested columns #105

Merged
merged 8 commits into from
Nov 28, 2023
Merged

Support bloom filters in lists/nested columns #105

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
57249d0
WIP debugging read of fruits.parquet
shannonwells Nov 21, 2023
d2c03c4
don't change the generated file location
shannonwells Nov 21, 2023
445f27e
all tests passing
shannonwells Nov 21, 2023
f76ae06
cleanup
shannonwells Nov 21, 2023
9957399
more cleanup
shannonwells Nov 21, 2023
b119c39
undo whitespace-only changes
shannonwells Nov 21, 2023
df9d247
write test files in bloomFilterIntegration.ts, add note to existing i…
shannonwells Nov 21, 2023
716c79d
Update lib/writer.ts
shannonwells Nov 22, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions lib/bloomFilterIO/bloomFilterReader.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,11 +117,11 @@ export const siftAllByteOffsets = (
};

export const getBloomFiltersFor = async (
columnNames: Array<string>,
paths: Array<string>,
envelopeReader: InstanceType<typeof ParquetEnvelopeReader>
) => {
const columnChunkDataCollection = envelopeReader.getAllColumnChunkDataFor(
columnNames
paths
);
const bloomFilterOffsetData = siftAllByteOffsets(columnChunkDataCollection);
const offsetByteValues = bloomFilterOffsetData.map(
Expand Down
40 changes: 28 additions & 12 deletions lib/reader.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,11 +6,25 @@ import * as parquet_schema from './schema';
import * as parquet_codec from './codec';
import * as parquet_compression from './compression';
import * as parquet_types from './types';
import BufferReader , { BufferReaderOptions } from './bufferReader';
import BufferReader, {BufferReaderOptions} from './bufferReader';
import * as bloomFilterReader from './bloomFilterIO/bloomFilterReader';
import fetch from 'cross-fetch';
import { ParquetCodec, Parameter,PageData, SchemaDefinition, ParquetType, FieldDefinition, ParquetField, ClientS3, ClientParameters, FileMetaDataExt, NewPageHeader, RowGroupExt, ColumnChunkExt } from './declare';
import { Cursor, Options } from './codec/types';
import {
ParquetCodec,
Parameter,
PageData,
SchemaDefinition,
ParquetType,
FieldDefinition,
ParquetField,
ClientS3,
ClientParameters,
FileMetaDataExt,
NewPageHeader,
RowGroupExt,
ColumnChunkExt
} from './declare';
import {Cursor, Options} from './codec/types';

const {
getBloomFiltersFor,
Expand All @@ -35,7 +49,7 @@ const PARQUET_RDLVL_ENCODING = 'RLE';
/**
* A parquet cursor is used to retrieve rows from a parquet file in order
*/
class ParquetCursor {
class ParquetCursor {

metadata: FileMetaDataExt;
envelopeReader: ParquetEnvelopeReader;
Expand All @@ -44,13 +58,14 @@ class ParquetCursor {
rowGroup: Array<unknown>;
rowGroupIndex: number;
cursorIndex: number;

/**
* Create a new parquet reader from the file metadata and an envelope reader.
* It is usually not recommended to call this constructor directly except for
* advanced and internal use cases. Consider using getCursor() on the
* ParquetReader instead
*/
constructor( metadata: FileMetaDataExt, envelopeReader: ParquetEnvelopeReader, schema: parquet_schema.ParquetSchema, columnList: Array<Array<unknown>>) {
constructor(metadata: FileMetaDataExt, envelopeReader: ParquetEnvelopeReader, schema: parquet_schema.ParquetSchema, columnList: Array<Array<unknown>>) {
this.metadata = metadata;
this.envelopeReader = envelopeReader;
this.schema = schema;
Expand All @@ -72,9 +87,9 @@ class ParquetCursor {
}

let rowBuffer = await this.envelopeReader.readRowGroup(
this.schema,
this.metadata.row_groups[this.rowGroupIndex],
this.columnList);
this.schema,
this.metadata.row_groups[this.rowGroupIndex],
this.columnList);

this.rowGroup = parquet_shredder.materializeRecords(this.schema, rowBuffer);
this.rowGroupIndex++;
Expand Down Expand Up @@ -363,7 +378,7 @@ export class ParquetReader {
}

decodePages(buffer: Buffer, opts: Options) {
return decodePages(buffer,opts);
return decodePages(buffer, opts);
}

}
Expand All @@ -375,6 +390,7 @@ export class ParquetReader {
* rows from a parquet file use the ParquetReader instead
*/
let ParquetEnvelopeReaderIdCounter = 0;

export class ParquetEnvelopeReader {
readFn: (offset: number, length: number, file?: string) => Promise<Buffer>;
close: () => unknown;
Expand All @@ -393,7 +409,7 @@ export class ParquetEnvelopeReader {
return Promise.reject('external references are not supported');
}

return parquet_util.fread(fileDescriptor, offset, length);
return parquet_util.fread(fileDescriptor, offset, length);
};

let closeFn = parquet_util.fclose.bind(undefined, fileDescriptor);
Expand All @@ -407,15 +423,15 @@ export class ParquetEnvelopeReader {
return Promise.reject('external references are not supported');
}

return Promise.resolve(buffer.slice(offset,offset+length));
return Promise.resolve(buffer.slice(offset, offset + length));
};

let closeFn = () => ({});
return new ParquetEnvelopeReader(readFn, closeFn, buffer.length, options);
}

static async openS3(client: ClientS3, params: ClientParameters, options?: BufferReaderOptions) {
let fileStat = async () => client.headObject(params).promise().then((d: {ContentLength: number}) => d.ContentLength);
let fileStat = async () => client.headObject(params).promise().then((d: { ContentLength: number }) => d.ContentLength);

let readFn = async (offset: number, length: number, file?: string) => {
if (file) {
Expand Down
Loading
Loading