Skip to content

AlexisJehan/DsvMender

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Logo

DSV Mender

maven-central javadoc build coverage license

A Java 11+ library to fix malformed DSV (Delimiter-Separated Values) data automatically.

Introduction

As many developers you may already have to treat some input data with formats such as CSV or JSON. Sometimes that task could become tricky to achieve because some values are not always formatted how they are supposed to be. DSV Mender is a library that aims to help you in such cases efficiently. Basically it collects some features from each valid column of the data independently to find the best solution while handling invalid or missing values.

Constraints and estimations

DSV Mender is working with a concept of constraints and estimations that are associated to specific columns of the data:

  • Constraints eliminate some candidate possibilities of a malformed row if they do not respect a rule, without taking into account previous valid values at all. For example if the third column has to be exactly 5 characters long, then all candidates with a value that does not will be discarded.

  • Estimations could be used to collect some features from valid values. When an invalid value needs to be fixed then the closest generated possibility is chosen. For example if you collect the length of valid values and get 5 characters 95% of the time then a possible fixed-value that got a length of 5 got more chances to be selected than a candidate of 3 characters.

Getting started

To include and use DSV Mender, you need to add the following dependency from the Maven Central:

Maven (pom.xml)

<dependency>
	<groupId>com.github.alexisjehan</groupId>
	<artifactId>dsv-mender</artifactId>
	<version>1.0.0</version>
</dependency>

Gradle Groovy (build.gradle)

plugins {
	id 'java-library'
}

repositories {
	mavenCentral()
}

dependencies {
	implementation 'com.github.alexisjehan:dsv-mender:1.0.0'
}

Gradle Kotlin (build.gradle.kts)

plugins {
	`java-library`
}

repositories {
	mavenCentral()
}

dependencies {
	implementation("com.github.alexisjehan:dsv-mender:1.0.0")
}

Also, the Javadoc can be accessed here.

Examples

Let's illustrate how it works step-by-step, consider the following CSV data:

Release,Release date,Highlights
Java SE 9,2017-09-21,Initial release
Java SE 9.0.1,2017-10-17,October 2017 security fixes and critical bug fixes
Java SE 9.0.4,2018-01-16,Final release for JDK 9; January 2018 security fixes and critical bug fixes
Java SE 10,2018-03-20,Initial release
Java SE 10.0.1,2018-04-17,Security fixes, 5 bug fixes
Java SE 11,2018-09-25,Initial release
Java SE 11.0.1,2018-10-16,Security & bug fixes
Java SE 11.0.2,2019-01-15,Security & bug fixes
Java SE 12,Initial release

As you may see, some lines are not well-formatted. The "Java SE 10.0.1" "Highlights" column contains the delimiter character, and the "Java SE 12" "Release date" column is missing. Let's see how to use DSV Mender to fix it.

Building the mender

First you need to create a Mender object based on the input data. That requires specifying the delimiter string as well as the expected number of columns.

Basic configuration

The lazy way, for a first attempt is to build a basic Mender, that can be able to mend most of the input data:

final var delimiter = ',';
final var length = 3;
final var mender = DsvMender.basic(delimiter, length);

Advanced configuration

For more accurate results, you can also build a Mender with custom Constraints and Estimations. For our example above we will use the following ones:

final var mender = DsvMender.builder()
		.withDelimiter(',')
		.withLength(3)
		.withConstraint(value -> value.startsWith("Java SE"), 0) // values[0] must start with "Java SE"
		.withConstraint(value -> value.isEmpty() || 10 == value.length(), 1)// values[1] must be empty or have a length of 10
		.build();

Processing the data

Once you get your Mender component built, you are able to process your data line by line. Note that you do not have to worry of the passed line being valid or not, if it is then the Mender will still fit its Estimations before returning it.

String row;
while (null != (row = reader.readLine())) {
	printValues(mender.mend(row));
}

Finally, here is the result we got for our example:

"Release", "Release date", "Highlights"
"Java SE 9", "2017-09-21", "Initial release"
"Java SE 9.0.1", "2017-10-17", "October 2017 security fixes and critical bug fixes"
"Java SE 9.0.4", "2018-01-16", "Final release for JDK 9; January 2018 security fixes and critical bug fixes"
"Java SE 10", "2018-03-20", "Initial release"
"Java SE 10.0.1", "2018-04-17", "Security fixes, 5 bug fixes"
"Java SE 11", "2018-09-25", "Initial release"
"Java SE 11.0.1", "2018-10-16", "Security & bug fixes"
"Java SE 11.0.2", "2019-01-15", "Security & bug fixes"
"Java SE 12", "", "Initial release"

(You can find the code of that example among others in the "examples" package)

Specific Maven goals

Generate the Javadoc API documentation:

mvn javadoc:javadoc

Generate the Jacoco test coverage report:

mvn jacoco:report

Update sources license:

mvn license:format

License

This project is licensed under the MIT License.