Java (embedded)

QuestDB is written in Java and can be used as any other Java library. Moreover, it is a single JAR with no additional dependencies.

To include QuestDB in your project, use the following:

Maven
Gradle

JDK11

<dependency>
  <groupId>org.questdb</groupId>
  <artifactId>questdb</artifactId>
  <version>6.5.2</version>
</dependency>
      

JDK8

<dependency>
  <groupId>org.questdb</groupId>
  <artifactId>questdb-jdk8</artifactId>
  <version>6.5.2</version>
</dependency>
      

Writing data#

The TableWriter facilitates table writes. To successfully create an instance of TableWriter, the table must:

Already exist
Have no other open writers against it as the TableWriter constructor will attempt to obtain an exclusive cross-process lock on the table.

Example table writer

final CairoConfiguration configuration = new DefaultCairoConfiguration("data_dir");
try (CairoEngine engine = new CairoEngine(configuration)) {
    final SqlExecutionContextImpl ctx = new SqlExecutionContextImpl(engine, 1);
    try (SqlCompiler compiler = new SqlCompiler(engine)) {

        compiler.compile("create table abc (a int, b byte, c short, d long, e float, g double, h date, i symbol, j string, k boolean, l geohash(8c), ts timestamp) timestamp(ts)", ctx);

        try (TableWriter writer = engine.getWriter(ctx.getCairoSecurityContext(), "abc", "testing")) {
            for (int i = 0; i < 11; i++) {
                TableWriter.Row row = writer.newRow(Os.currentTimeMicros());
                row.putInt(0, 123);
                row.putByte(1, (byte) 1111);
                row.putShort(2, (short) 222);
                row.putLong(3, 333);
                row.putFloat(4, 4.44f);
                row.putDouble(5, 5.55);
                row.putDate(6, System.currentTimeMillis());
                row.putSym(7, "xyz");
                row.putStr(8, "abc");
                row.putBool(9, true);
                row.putGeoHash(10, GeoHashes.fromString("u33dr01d", 0, 8));
                row.append();
            }
            writer.commit();
        }
    }
}

Detailed steps#

Configure Cairo engine#

CairoEngine is a resource manager for the embedded QuestDB. Its main function is to facilitate concurrent access to pools of TableReader and TableWriter instances.

New CairoEngine instance

final CairoConfiguration configuration = new DefaultCairoConfiguration("data_dir");
try (CairoEngine engine = new CairoEngine(configuration)) {

A typical application will need only one instance of CairoEngine. This instance will start when application starts and shuts down when application closes. You will need to close CairoEngine gracefully when the application stops.

QuestDB provides a default configuration which only requires the data directory to be specified. For a more advanced usage, the whole CairoConfiguration interface can be overridden.

Create an instance of SqlExecutionContext#

Execution context is a conduit for passing SQL execution artefacts to the execution site. This instance is not thread-safe and it must not be shared between threads.

Example of execution context

final SqlExecutionContextImpl ctx = new SqlExecutionContextImpl(engine, 1);

The second argument is the number of threads that will be helping to execute SQL statements. Unless you are building another QuestDB server, this value should always be 1.

New SqlCompiler instance and blank table#

Before we start writing data using TableWriter, the target table has to exist. There are several ways to create new table ; using SqlCompiler is the easiest.

Creating new table

try (SqlCompiler compiler = new SqlCompiler(engine)) {
    compiler.compile("create table abc (a int, b byte, c short, d long, e float, g double, h date, i symbol, j string, k boolean, l geohash(8c), ts timestamp) timestamp(ts)", ctx);

As you will be able to see below, the table field types and indexes must match the code that is populating the table.

New instance of TableWriter#

We use engine to create instance of TableWriter. This will enable reusing this TableWriter instance later, when we use the same method of creating table writer again.

New table writer instance

try (TableWriter writer = engine.getWriter(ctx.getCairoSecurityContext(), "abc", "testing")) {

The writer will hold exclusive lock on table abc until it is closed and testing will be used as the lock reason. This lock is both intra and inter-process. If you have two Java applications accessing the same table only one will succeed at one time.

Create a new row#

Creating new table row with timestamp

TableWriter.Row row = writer.newRow(Os.currentTimeMicros());

Although this operation semantically looks like a new object creation, the row instance is actually being re-used under the hood. A Timestamp is necessary to determine a partition for the new row. Its value has to be either increment or stay the same as the last row. When the table is not partitioned and does not have a designated timestamp column, timestamp value can be omitted.

Creating new table row without timestamp

TableWriter.Row row = writer.newRow();

Populate columns#

There are put* methods for every supported data type. Columns are updated by an index as opposed to by name.

Populating table column

row.putLong(3, 333);

Column update order is not important and update can be sparse. All unset columns will default to NULL values.

Append a row#

Following method call:

Appending a new row

row.append();

Appended rows are not visible to readers until they are committed. An unneeded row can also be canceled if required.

Cancelling half-populated row

row.cancel();

A pending row is automatically cancelled when writer.newRow() is called. Consider the following scenario:

TableWriter.Row row = writer.newRow(Os.currentTimeMicros());
row.putInt(0, 123);
row.putByte(1, (byte) 1111);
row.putShort(2, (short) 222);
row.putLong(3, 333);
row = writer.newRow(Os.currentTimeMicros());
...

Second newRow() call would cancel all the updates to the row since the last append().

Commit changes#

To make changes visible to readers, writer has to commit. writer.commit does this job. Unlike traditional SQL databases, the size of the transaction does not matter. You can commit anything between 1 and 1 trillion rows. We also spent considerable effort to ensure commit() is lightweight. You can drip one row at a time in applications that require such behaviour.

Writing columns in blocks#

QuestDB supports writing blocks of columnar data at once via the use of the TableBlockWriter. The TableBlockWriter instance is obtained from a TableWriter and can then be used to write in memory frames of columnar data. A frame of columnar data is just a piece of contiguous memory with each column value stored in it one after another. The TableBlockWriter will allow any number of such frames of columnar data to be written with an invocation of the appendPageFrameColumn method, before the block is either committed or cancelled (rolled back). Use of the TableBlockWriter requires that all columns have the same number of rows written to them and within each column the frames need to be added in append order.

A PageFrame instance can optionally be used as a convenient interface to hold the columnar frames and a PageFrameCursor instance can be used as an interface to provide a sequence of frames to be committed. Many of QuestDB's RecordCursorFactory implementations provide a PageFrameCursor.

Example table block writer

final CairoConfiguration configuration = new DefaultCairoConfiguration("data_dir");
try (CairoEngine engine = new CairoEngine(configuration)) {
    final SqlExecutionContextImpl ctx = new SqlExecutionContextImpl(engine, 1);
    try (SqlCompiler compiler = new SqlCompiler(engine)) {

        PageFrameCursor cursor = ...; // Setup PageFrameCursor instance
        compiler.compile("create table abc (a int, b byte, c short, d long, e float, g double, h date, i symbol, j string, k boolean, l geohash(8c), ts timestamp) timestamp(ts)", ctx);

        try (TableWriter writer = engine.getWriter(ctx.getCairoSecurityContext(), "abc", "testing")) {
            int columnCount = writer.getMetadata().getColumnCount();
            TableBlockWriter blockWriter = writer.newBlock();

            PageFrame frame;
            while ((frame = cursor.next()) != null) {
                for (int columnIndex = 0; columnIndex < columnCount; columnIndex++) {
                    blockWriter.appendPageFrameColumn(
                            columnIndex,
                            frame.getPageSize(columnIndex),
                            frame.getPageAddress(columnIndex));
                }
            }
            blockWriter.commit();
        }
    }
}

Executing queries#

We provide a single API for executing all kinds of SQL queries. The example below focuses on SELECT and how to fetch data from a cursor.

Compiling SQL

final CairoConfiguration configuration = new DefaultCairoConfiguration(temp.getRoot().getAbsolutePath());
try (CairoEngine engine = new CairoEngine(configuration)) {
    final SqlExecutionContextImpl ctx = new SqlExecutionContextImpl(engine, 1);
    try (SqlCompiler compiler = new SqlCompiler(engine)) {
        try (RecordCursorFactory factory = compiler.compile("abc", ctx).getRecordCursorFactory()) {
            try (RecordCursor cursor = factory.getCursor(ctx)) {
                final Record record = cursor.getRecord();
                while (cursor.hasNext()) {
                    // access 'record' instance for field values
                }
            }
        }
    }
}

Detailed steps#

The steps to setup CairoEngine, execution context and SqlCompiler are the same as those we explained in writing data section. We will skip them here and focus on fetching data.

RecordCursorFactory#

You can think of RecordCursorFactory as PreparedStatement. This is the entity that holds SQL execution plan with all of the execution artefacts. Factories are designed to be reused and we strongly encourage caching them. You also need to make sure that you close factories explicitly when you no longer need them. Failing to do so can cause memory and/or other resources leak.

RecordCursor#

This instance allows iterating over the dataset produced by SQL. Cursors are relatively short-lived and do not imply fetching all the data. Note that you have to close a cursor as soon as enough data is fetched ; the closing process can happen at any time.

Cursors are not thread safe and cannot be shared between threads.

Record#

This is cursor's data access API. Record instance is obtained from the cursor outside of the fetch loop.

Example of fetching data from cursor

final Record record = cursor.getRecord();
while (cursor.hasNext()) {
    // access 'record' instance for field values
}

Record does not hold the data. Instead, it is an API to pull data when data is needed. Record instance remains the same while cursor goes over the data, making caching of records pointless.

InfluxDB sender library#

QuestDB library provides fast and efficient way of sending line protocol messages. Sender implementation entry point is io.questdb.cutlass.line.udp.LineUdpSender, it is fully zero-GC and has latency in a region of 200ns per message.

Get started#

Step 1: Create an instance of LineUdpSender.

Arguments	Description
`interfaceIPv4Address`	Network interface to use to send messages.
`sendToIPv4Address`	Destination IP address
`bufferCapacity`	Send buffer capacity to batch messages. Do not configure this buffer over the MTU size
`ttl`	UDP packet TTL. Set this number appropriate to how many VLANs your messages have to traverse before reaching the destination

Example

LineUdpSender sender = new LineUdpSender(0, Net.parseIPv4("232.1.2.3"), 9009, 110, 2);

Step 2: Create entries by building each entry's tagset and fieldset.

Syntax

sender.metric("table_name").tag("key","value").field("key", value).$(timestamp);

where

Element	Description	Can be repeated
`metric(tableName)`	Specify which table the data is to be written into	no
`tag("key","value")`	Use to add a new key-value entry as metadata	yes
`field("key","value")`	Use to add a new key-value entry as reading	yes
`$(timestamp)`	Specify the timestamp for the reading	no

tip

You can chain several tags and fields, e.g tag("a","x").tag("b","y").tag("c","z") etc.

Example

sender.metric("readings").tag("city", "London").tag("by", "quest").field("temp", 3400).field("humid", 0.434).$(Os.currentTimeNanos());

Sender will send message as soon as send buffer is full. To send messages before buffer fills up you can use sender.flush()

Full example#

This example sends multicast messages to 232.1.2.3:9009. In this mode multiple QuestDB instances can receive the same message.

Sending InfluxDB line protocol

LineUdpSender sender = new LineUdpSender(0, Net.parseIPv4("232.1.2.3"), 9009, 1024, 2);
sender.metric("readings").tag("city", "London").tag("by", "quest").field("temp", 3400).$(Os.currentTimeMicros());
sender.metric("readings").tag("city", "London").tag("by", "quest").field("temp", 3400).$(Os.currentTimeMicros());
sender.metric("readings").tag("city", "London").tag("by", "quest").field("temp", 3400).$(Os.currentTimeMicros());
sender.flush();