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.

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

DRILL-8474: Add Daffodil Format Plugin #2836

Closed
wants to merge 8 commits into from
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
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
27 changes: 27 additions & 0 deletions contrib/format-daffodil/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
# Daffodil 'Format' Reader
This plugin enables Drill to read DFDL-described data from files by way of the Apache Daffodil DFDL implementation.

## Validation

Data read by Daffodil is always validated using Daffodil's Limited Validation mode.

TBD: do we need an option to control escalating validation errors to fatal? Currently this is not provided.

## Limitations: TBD

At the moment, the DFDL schema is found on the local file system, which won't support Drill's distributed architecture.

There are restrictions on the DFDL schemas that this can handle.

In particular, all element children must have distinct element names, including across choice branches.
(This rules out a number of large DFDL schemas.)

TBD: Auto renaming as part of the Daffodil-to-Drill metadata mapping?

The data is parsed fully from its native form into a Drill data structure held in memory.
No attempt is made to avoid access to parts of the DFDL-described data that are not needed to answer the query.

If the data is not well-formed, an error occurs and the query fails.

If the data is invalid, and validity checking by Daffodil is enabled, then an error occurs and the query fails.

94 changes: 94 additions & 0 deletions contrib/format-daffodil/pom.xml
Original file line number Diff line number Diff line change
@@ -0,0 +1,94 @@
<?xml version="1.0"?>
<!--

Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements. See the NOTICE file
distributed with this work for additional information
regarding copyright ownership. The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

-->
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>

<parent>
<artifactId>drill-contrib-parent</artifactId>
<groupId>org.apache.drill.contrib</groupId>
<version>1.22.0-SNAPSHOT</version>
</parent>

<artifactId>drill-format-daffodil</artifactId>
<name>Drill : Contrib : Format : Daffodil</name>

<dependencies>
<dependency>
<groupId>org.apache.drill.exec</groupId>
<artifactId>drill-java-exec</artifactId>
<version>${project.version}</version>
</dependency>
<dependency>
<groupId>org.apache.daffodil</groupId>
<artifactId>daffodil-japi_2.12</artifactId>
<version>3.7.0</version>
</dependency>
<dependency>
<groupId>org.apache.daffodil</groupId>
<artifactId>daffodil-runtime1_2.12</artifactId>
<version>3.7.0</version>
</dependency>
<!-- Test dependencies -->
<dependency>
<groupId>org.apache.drill.exec</groupId>
<artifactId>drill-java-exec</artifactId>
<classifier>tests</classifier>
<version>${project.version}</version>
<scope>test</scope>
</dependency>

<dependency>
<groupId>org.apache.drill</groupId>
<artifactId>drill-common</artifactId>
<classifier>tests</classifier>
<version>${project.version}</version>
<scope>test</scope>
</dependency>
</dependencies>

<build>
<plugins>
<plugin>
<artifactId>maven-resources-plugin</artifactId>
<executions>
<execution>
<id>copy-java-sources</id>
<phase>process-sources</phase>
<goals>
<goal>copy-resources</goal>
</goals>
<configuration>
<outputDirectory>${basedir}/target/classes/org/apache/drill/exec/store/daffodil
</outputDirectory>
<resources>
<resource>
<directory>src/main/java/org/apache/drill/exec/store/daffodil</directory>
<filtering>true</filtering>
</resource>
</resources>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build>
</project>
Original file line number Diff line number Diff line change
@@ -0,0 +1,181 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

package org.apache.drill.exec.store.daffodil;

import org.apache.daffodil.japi.DataProcessor;
import org.apache.drill.common.AutoCloseables;
import org.apache.drill.common.exceptions.CustomErrorContext;
import org.apache.drill.common.exceptions.UserException;
import org.apache.drill.exec.physical.impl.scan.v3.ManagedReader;
import org.apache.drill.exec.physical.impl.scan.v3.file.FileDescrip;
import org.apache.drill.exec.physical.impl.scan.v3.file.FileSchemaNegotiator;
import org.apache.drill.exec.physical.resultSet.RowSetLoader;
import org.apache.drill.exec.record.metadata.TupleMetadata;
import org.apache.drill.exec.store.daffodil.schema.DaffodilDataProcessorFactory;
import org.apache.drill.exec.store.dfs.DrillFileSystem;
import org.apache.drill.exec.store.dfs.easy.EasySubScan;
import org.apache.hadoop.fs.Path;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.IOException;
import java.io.InputStream;
import java.net.URI;
import java.net.URISyntaxException;
import java.util.Objects;

import static org.apache.drill.exec.store.daffodil.schema.DaffodilDataProcessorFactory.*;
import static org.apache.drill.exec.store.daffodil.schema.DrillDaffodilSchemaUtils.daffodilDataProcessorToDrillSchema;

public class DaffodilBatchReader implements ManagedReader {

private static final Logger logger = LoggerFactory.getLogger(DaffodilBatchReader.class);
private final RowSetLoader rowSetLoader;
private final CustomErrorContext errorContext;
private final DaffodilMessageParser dafParser;
private final InputStream dataInputStream;

public DaffodilBatchReader(DaffodilReaderConfig readerConfig, EasySubScan scan,
FileSchemaNegotiator negotiator) {

errorContext = negotiator.parentErrorContext();
DaffodilFormatConfig dafConfig = readerConfig.plugin.getConfig();

String schemaURIString = dafConfig.getSchemaURI(); // "schema/complexArray1.dfdl.xsd";
String rootName = dafConfig.getRootName();
String rootNamespace = dafConfig.getRootNamespace();
boolean validationMode = dafConfig.getValidationMode();

URI dfdlSchemaURI;
try {
mbeckerle marked this conversation as resolved.
Show resolved Hide resolved
dfdlSchemaURI = new URI(schemaURIString);
} catch (URISyntaxException e) {
throw UserException.validationError(e).build(logger);
}

FileDescrip file = negotiator.file();
DrillFileSystem fs = file.fileSystem();
URI fsSchemaURI = fs.getUri().resolve(dfdlSchemaURI);

DaffodilDataProcessorFactory dpf = new DaffodilDataProcessorFactory();
DataProcessor dp;
try {
dp = dpf.getDataProcessor(fsSchemaURI, validationMode, rootName, rootNamespace);
} catch (CompileFailure e) {
throw UserException.dataReadError(e)
.message(String.format("Failed to get Daffodil DFDL processor for: %s", fsSchemaURI))
.addContext(errorContext).addContext(e.getMessage()).build(logger);
}
// Create the corresponding Drill schema.
// Note: this could be a very large schema. Think of a large complex RDBMS schema,
// all of it, hundreds of tables, but all part of the same metadata tree.
TupleMetadata drillSchema = daffodilDataProcessorToDrillSchema(dp);
// Inform Drill about the schema
negotiator.tableSchema(drillSchema, true);

//
// DATA TIME: Next we construct the runtime objects, and open files.
//
// We get the DaffodilMessageParser, which is a stateful driver for daffodil that
// actually does the parsing.
rowSetLoader = negotiator.build().writer();

// We construct the Daffodil InfosetOutputter which the daffodil parser uses to
// convert infoset event calls to fill in a Drill row via a rowSetLoader.
DaffodilDrillInfosetOutputter outputter = new DaffodilDrillInfosetOutputter(rowSetLoader);

mbeckerle marked this conversation as resolved.
Show resolved Hide resolved
// Now we can set up the dafParser with the outputter it will drive with
// the parser-produced infoset.
dafParser = new DaffodilMessageParser(dp); // needs further initialization after this.
dafParser.setInfosetOutputter(outputter);

Path dataPath = file.split().getPath();
// Lastly, we open the data stream
try {
dataInputStream = fs.openPossiblyCompressedStream(dataPath);
} catch (IOException e) {
throw UserException.dataReadError(e)
.message(String.format("Failed to open input file: %s", dataPath.toString()))
.addContext(errorContext).addContext(e.getMessage()).build(logger);
}
// And lastly,... tell daffodil the input data stream.
dafParser.setInputStream(dataInputStream);
}

/**
* This is the core of actual processing - data movement from Daffodil to Drill.
* <p>
* If there is space in the batch, and there is data available to parse then this calls the
* daffodil parser, which parses data, delivering it to the rowWriter by way of the infoset
* outputter.
* <p>
* Repeats until the rowWriter is full (a batch is full), or there is no more data, or a parse
* error ends execution with a throw.
* <p>
* Validation errors and other warnings are not errors and are logged but do not cause parsing to
* fail/throw.
*
* @return true if there are rows retrieved, false if no rows were retrieved, which means no more
* will ever be retrieved (end of data).
* @throws RuntimeException
* on parse errors.
*/
@Override
public boolean next() {
// Check assumed invariants
// We don't know if there is data or not. This could be called on an empty data file.
// We DO know that this won't be called if there is no space in the batch for even 1
// row.
if (dafParser.isEOF()) {
return false; // return without even checking for more rows or trying to parse.
}
while (rowSetLoader.start() && !dafParser.isEOF()) { // we never zero-trip this loop.
// the predicate is always true once.
dafParser.parse();
if (dafParser.isProcessingError()) {
assert (Objects.nonNull(dafParser.getDiagnostics()));
throw UserException.dataReadError().message(dafParser.getDiagnosticsAsString())
.addContext(errorContext).build(logger);
}
if (dafParser.isValidationError()) {
logger.warn(dafParser.getDiagnosticsAsString());
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need an option here to convert validation errors to fatal?

Will logger.warn be seen by a query user, or is that just for someone dealing with the logs?

Validation errors either should be escalated to fatal, OR they should be visible in the query output display to a user somehow.

Either way, users will need a mechanism to suppress validation errors that prove to be unavoidable since they could be common place. Nodody wants thousands of warnings about something they can't avoid that doesn't stop parsing and querying the data.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mbeckerle The question I'd have is whether the query can proceed if validation fails. (I don't know the answer)
If the answer is no, then we need to halt execution ASAP and throw an exception. If the answer is it can proceed, but the data might be less than ideal, maybe we add a configuration option which will allow the user to decide the behavior on a validation failure.

I could imagine situations where you have Drill unable to read a huge file because someone fat fingered a quotation mark somewhere or something like that. In a situation like that, sometimes you might just want to say I'll accept a row or two of bad data just so I can read the whole file.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agree.

We draw a distinction between "well formed" and "invalid" data and whether one does validation seems like the right switch in daffodil to use.

If data is malformed, that means you can't successfully parse it. If it is invalid, that just means values are unexpected. Example: A 3 digit number representing a percentage 0 to 100. -1 is invalid, ABC is malformed.

If data is not well formed, you really cannot continue parsing it, as you cannot convert it to the type expected. But, if you are able to determine at least how big it is, it's possible to capture that length of data into a dummy "badData" element which is always invalid (so isn't a "false positive" parse). This capability has to be designed into the DFDL schema, but it is something we've been doing more and more.

Hence, one can tolerate even some malformed data. If it is malformed to where you cannot determine the length, then continuing is impossible.

We will see if more than this is needed. Options like the "use all strings/varchar" or all numbers are float, which you have for toleratng situations with other data connectors may prove useful, particularly while a DFDL schema is in development and you are really just testing it (and the corresponding data) using Drill.

// Note that even if daffodil is set to not validate, validation errors may still occur
// from DFDL's "recoverableError" assertions.
}
rowSetLoader.save();
}
int nRows = rowSetLoader.rowCount();
mbeckerle marked this conversation as resolved.
Show resolved Hide resolved
assert nRows > 0; // This cannot be zero. If the parse failed we will have already thrown out
// of here.
return true;
}

@Override
public void close() {
AutoCloseables.closeSilently(dataInputStream);
cgivre marked this conversation as resolved.
Show resolved Hide resolved
}
}

class DaffodilReaderConfig {
final DaffodilFormatPlugin plugin;

DaffodilReaderConfig(DaffodilFormatPlugin plugin) {
this.plugin = plugin;
}
}
Loading