The MarkLogic Java API support reading, writing, and deleting document content and metadata as well as querying documents.
Before using the API, create a REST server for your database. If the database has existing documents, modify the roles with privileges to read and write those documents to inherit from rest-reader and rest-writer.
Use {@link com.marklogic.client.DatabaseClientFactory}.newClient() to create a client for a database. You can then use the {@link com.marklogic.client.DatabaseClient} factory methods (which begin with "new") to create document, configuration, and query managers.
Use the {@link com.marklogic.client.DatabaseClient} to create the appropriate document manager for the format of the content -- a {@link com.marklogic.client.document.BinaryDocumentManager} for a binary document or an {@link com.marklogic.client.document.XMLDocumentManager} for an XML document. The client also provides {@link com.marklogic.client.document.JSONDocumentManager} and {@link com.marklogic.client.document.TextDocumentManager} document manager for those formats. Use {@link com.marklogic.client.document.GenericDocumentManager} if you don't know the format or need to process multiple documents with different formats.
To identify a document in the database, you use a String with the document URI.
To access document content, you can create a handle object that provides strongly-typed support for an IO representation. This concept of content handles is core to the Java API (making use of the Adapter design pattern so the API can read or write content with a diverse and extensible set of representations). You can define new handle classes to integrate new IO representations into the API.
For frequently-used requests, you can also write an IO object or specify an IO class to read an object directly. Internally, these shortcut methods create the handle for you. By default, the IO representations supported in shortcut methods include byte[], DOM Document, File, InputStream, SAX InputSource, Reader, Transformer Source, String, StAX XMLEventReader, and StAX XMLEventWriter. You can enable shortcut support for dom4j Document, JDOM Document, GSON JsonElement, Jackson JsonNode, or JAXB POJOs.
For instance, consider a binary JPEG file. To write this file to the database using a strongly typed handle, you create a {@link com.marklogic.client.io.FileHandle} and call the handle's set() method with the File object. You then pass the handle to the operation that writes the binary content to the database.
Similarly, let's say you want to read the content of an XML document from the database into a DOM document. You pass a {@link com.marklogic.client.io.DOMHandle} to the operation that reads the XML content. The read operation populates the DOMHandle with the content. You can then call the handle's get() method to get the org.w3c.dom.Document object. In short, a handle acts as a container for content in a representation.
To read document content from the database, pass the document URI string
and a handle to the read() method
of the document manager. Then use the methods of the handle to get the content.
The following example reads the content of the /db/path/to/myDoc.xml
database document into a DOM document:
XMLDocumentManager docMgr = client.newXMLDocumentManager(); String docId = "/db/path/to/myDoc.xml"; DOMHandle handle = new DOMHandle(); docMgr.read(docId, handle); Document document = handle.get();
The following example reads the document with the equivalent shortcut method:
XMLDocumentManager docMgr = client.newXMLDocumentManager(); String docId = "/db/path/to/myDoc.xml"; Document document = docMgr.readAs(docId, Document.class);
To write content to a document in the database, set the content in a handle and then pass the
document URI string and the content handle to the write() method
of the document manager. The following example write the content of the
/db/path/to/myDoc.xml database document from a DOM document:
Document document = ... create or modify the document ...; XMLDocumentManager docMgr = client.newXMLDocumentManager(); String docId = "/db/path/to/myDoc.xml"; DOMHandle handle = new DOMHandle(); handle.set(document); docMgr.write(docId, handle);
The following example writes the document with the equivalent shortcut method:
Document document = ... create or modify the document ...; XMLDocumentManager docMgr = client.newXMLDocumentManager(); String docId = "/db/path/to/myDoc.xml"; docMgr.writeAs(docId, document);
To learn which handles you can use to read the content of an XML document, look at the classes that implement {@link com.marklogic.client.io.marker.XMLReadHandle}. To learn which handles you can use to write the content of an XML document, look at the classes that implement {@link com.marklogic.client.io.marker.XMLWriteHandle}. {@link com.marklogic.client.document.BinaryDocumentManager}, {@link com.marklogic.client.document.JSONDocumentManager}, {@link com.marklogic.client.document.TextDocumentManager}, and {@link com.marklogic.client.document.GenericDocumentManager} each have corresponding read and write handles that identify the handle classes that can be used when reading and writing documents of that format. The shortcut methods of each document manager work with the same IO objects supported by the handles that are accepted by the document manager and that implement {@link com.marklogic.client.io.marker.ContentHandle}.
You can read or write document metadata in the same request as document content or separately.
To prepare a database for search applications, you will have created indexes on the fields and elements of documents. Query options declare and configure the use of indexes for search and query applications. You specify query options as a JSON or XML structure. You use {@link com.marklogic.client.admin.QueryOptionsManager} to write query options in JSON or XML to the server or to read query options back into the application.
The following example writes new query options named "shipments" with a single constraint. The constraint, of type "value" names the "port" location within documents for queries that use the "shipments" query options.
QueryOptionsManager optionsMgr = client.newQueryOptionsManager();
String options = new StringBuilder()
.append("<search:options>")
.append( "<search:constraint name=\"port\">")
.append( "<search:value>")
.append( "<search:element name=\"port\" ns=\"\"/>")
.append( "</search:value>")
.append( "</search:constraint>")
.append("</search:options>")
.toString();
StringHandle handle = new StringHandle(options);
optionsMgr.writeOptions("shipments", handle);
The following example writes the query options with the equivalent shortcut method:
QueryOptionsManager optionsMgr = client.newQueryOptionsManager();
String options = new StringBuilder()
.append("<search:options>")
.append( "<search:constraint name=\"port\">")
.append( "<search:value>")
.append( "<search:element name=\"port\" ns=\"\"/>")
.append( "</search:value>")
.append( "</search:constraint>")
.append("</search:options>")
.toString();
optionsMgr.writeOptionsAs("shipments", Format.XML, options);
Use {@link com.marklogic.client.query.QueryManager} to search documents or extract values, co-occurrences of values, or full tuple records from document indexes. The search() method takes the name for search options that you wrote to the database previously, a query definition including the search criteria, and a handle for representing the results. You can supply search criteria as a string, keys-value pairs, or a structure. You can represent results as JSON, XML, or a Java data structure.
The following example uses the "shipments" query options written to database previously.
The search criteria matches documents that contain the term "electronics" as well as
a <port> element with a value of "Lisbon". The example reads
the search results into a DOM document.
QueryManager queryMgr = client.newQueryManager();
StringQueryDefinition querydef = queryMgr.newStringDefinition("shipments");
querydef.setCriteria("electronics port:Lisbon");
DOMHandle handle = new DOMHandle();
queryMgr.search(querydef, handle);
Document results = handle.get();
Besides XML or JSON, you can also get search results as a Java data structure by using {@link com.marklogic.client.io.SearchHandle}. This approach is often the most convenient way to manipulate search results as demonstrated in the following example:
QueryManager queryMgr = client.newQueryManager();
StringQueryDefinition querydef = queryMgr.newStringDefinition("shipments");
querydef.setCriteria("electronics port:Lisbon");
SearchHandle results = new SearchHandle();
queryMgr.search(querydef, results);
... do something at the response level (total response, metrics, and so on) ...
for (MatchDocumentSummary docSummary: results.getMatchResults()) {
... do something for a matching document ...
for (MatchLocation location : docSummary.getMatchLocations()) {
... do something at the match location level ...
for (MatchSnippet snippet : location.getSnippets()) {
if (snippet.isHighlighted()) {
... do something with highlighted text ...
} else {
... do something with unhighlighted text ...
}
}
}
}
The API makes use of the Fluent Interface design pattern so that, when you become familiar with the API, you can execute operations concisely.
For instance, because the read() method returns the content handle, the earlier read operation can be expressed in a single line:
org.w3c.dom.Document document =
client.newXMLDocumentManager().read(
client.newDocId("/db/path/to/myDoc.xml"),
new DOMHandle()
).get();
Because most handles provide a with() method that sets the content and
returns the handle, you can a write operation in a single line:
Document document = ... create or modify the document ...;
client.newXMLDocumentManager().write(
client.newDocId("/db/path/to/myDoc.xml"),
new DOMHandle().with(document)
);
When convenient, you can write a search operation in a single line:
Document results = client.newQueryManager().search(
queryMgr.newStringDefinition("shipments").withCriteria("electronics port:Lisbon"),
new DOMHandle()
).get();
To use extra IO handles like {@link com.marklogic.client.extra.dom4j.DOM4JHandle}, {@link com.marklogic.client.extra.gson.GSONHandle}, or {@link com.marklogic.client.extra.jdom.JDOMHandle}, you download the library integrated by the handle and add the library to the classpath. You can enable support in shortcut methods for the IO representation provided by an extra library by registering a factory for the handle using {@link com.marklogic.client.DatabaseClientFactory}.getHandleRegistry().
You can also register a factory for the built-in JAXB and JacksonDatabind handles to support IO on your POJO classes in shortcut methods. The following example registers the JAXB handle for one or more POJO classes and writes and reads the POJOs as database documents. The same pattern works with JacksonDatabindHandle.
// configure once before creating a client DatabaseClientFactory.getHandleRegistry().register( JAXBHandle.newFactory(Product.class, ... other POJO classes ...) ); // ... create a database client ... XMLDocumentManager docMgr = client.newXMLDocumentManager(); String docId = "/db/path/to/myPOJO.xml"; Product product = ... create or modify the POJO ...; docMgr.writeAs(docId, product); // ... at some other time ... Product product = docMgr.readAs(docId, Product.class);
We use slf4j for logging. This means you can choose any slf4j-compliant logging framework such as Logback, AVSL, JDK logging, Log4j, or Simple. If you don't know which to choose, we recommend Logback since it is a native implementation and the easiest to configure. Please follow the instructions on the slf4j website to configure your logging framework. It should take no more than 15 minutes. Once your logging framework is configured with slf4j, you should be able to see and manage logging from the Java-client API. This is especially important for long-running QueryBatcher and WriteBatcher jobs.