Comments (4)
Hi @isharamet ,
It looks like a bug. There is a Github issue created to replace the custom in
predicate with one recently introduced to Parquet: #272.
from parquet4s.
Fix to be released in 2.11.0
from parquet4s.
And I believe the error is in inverseCanDrop
method of InPredicate
:
override def inverseCanDrop(statistics: Statistics[T]): Boolean = {
val compare = statistics.getComparator.compare(_, _)
val min = statistics.getMin
val max = statistics.getMax
val isInRange = (value: T) => compare(value, min) >= 0 && compare(value, max) <= 0
values.exists(isInRange)
}
While it works for canDrop
, in not in
scenario it'll drop all the blocks with values from the set, even if blocks might contain other values. So, for example from my original post all values will be stored in a single block (min = 1
, max = 5
), so this block will be skipped.
from parquet4s.
Hi @isharamet,
We faced this bug too and ended up implementing our own NinPredicate similar to https://github.com/mjakubowski84/parquet4s/blob/master/core/src/main/scala/com/github/mjakubowski84/parquet4s/Filter.scala#L323 with both canDrop
and inverseCanDrop
returning false
. I don't see an option to utilize this "drop whole block if values not in range" logic, since for nin
it should rather be "accept whole block if values not in range".
This can be a temporary solution until nin
predicate is supported.
from parquet4s.
Related Issues (20)
- Reading from gcs bucket HOT 1
- Do not publish a pekko/akko versions of scapapb module HOT 1
- missing tail records of large(~193M) parquet files HOT 4
- Protobuf enums deserialisation HOT 3
- compatible parquet-hadoop with spark3.1 HOT 3
- Unsure how to use for 'semiauto' approach HOT 2
- ParquetSchemaResolver test fails on recent JVMs HOT 1
- [akka/pekko] Too many paths created during record partitioning HOT 2
- [RFC] Refactor timestamp codecs HOT 2
- Feature request: Expose partitions as a `Stream[F, Stream[F, Record]]` for FS2 HOT 5
- Incorrect value after reading parquet HOT 7
- [Question] get a listing of parquet files? HOT 4
- [Question] Is there a mechanism to detect when the `rotatingWriter` finishes writing to a file and to be notified of the file that was written? HOT 1
- Support vectored io introduced in Parquet 1.14 HOT 1
- Efficent way to read big files? HOT 2
- Possiblity to write avro IndexedRecords to Parquet using ParquetStreams HOT 4
- feat(akkaPekko): Retry mechanism for ParquetPartitioningFlow HOT 4
- Partitions with nested directories return zero rows HOT 1
- partitioning incompatibility with spark HOT 3
- ProjectionSchema self-inconsistency with partitioned source HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from parquet4s.