apache / apache/parquet-java

Filter predicate should be able to filter on columns not specified in projection pushdown

Aperta
#1,438 2 commenti 0 reazioni 0 assegnatari Vedi su GitHub
Component: Java Component: Parquet Priority: Minor Type: enhancement
Lingua principale
Java
Stelle
3.1k
Fork
1.6k
Merge medio
3g 12h
PR unite (30g)
33

Descrizione

When using both filter predicate and projection pushdown of columns, the column specified in the filter predicate, should not have to be specified in the projection pushdown.

For example, if I want the first names of students that are over the age of 16, then I should be able to specify ParquetInputFormat.setFilterPredicate(jobConf, IntColumn("age") > 16) and only have to specify ThriftReadSupport.setProjectionPushdown(jobConf, "fname")

**Reporter**: [Tim Chan](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=zzztimbo) / @zzztimbo

**Note**: *This issue was originally created as [PARQUET-295](https://issues.apache.org/jira/browse/PARQUET-295). Please see the [migration documentation](https://issues.apache.org/jira/browse/PARQUET-2502) for further details.*

Guida per i contributori

Nessuna guida per i contributori indicizzata per questo repository

Direzione di ricerca

Start with ParquetInputFormat.setFilterPredicate and ThriftReadSupport.setProjectionPushdown, using the example in the issue to trace how projected columns and filter columns are handled together. The work is done when a projection containing only fname can still filter on age > 16, with coverage for that behavior added to the relevant existing tests.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Valutazione

Stack tecnologico
java
Ambito
data
Tipo di issue
Funzionalità
Difficoltà
4/5
Tempo stimato
3-5 giorni
Stato di attività
Ferma
Chiarezza
Abbastanza chiara
Idoneità per principianti
38/100

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.