Android schema utils


0.7 (Feb 27, 2015)
Oct 8, 2013
May 4, 2020
Zendesk (zendesk)
Jerzy Chałupski (chalup)
Bartek Filipowicz (bartoszfilipowicz)
Lukáš Doležal (DocX)
Source code


Android schema utils

Android library for simplifying database schema and migrations management.


The library provides the fluent API, which allows you to define current schema:

private static final Schemas SCHEMA = Schemas.Builder
        new TableDefinition(Tables.ADDRESSES, ImmutableList.<TableDefinitionOperation>builder()
                new AddColumn(Addresses.REGION, "TEXT"),
                new AddColumn(Addresses.ZIP, "TEXT"),
                new AddColumn(Addresses.COUNTRY, "TEXT"),
                new AddColumn(Addresses.CITY, "TEXT"),
                new AddColumn(Addresses.STREET, "TEXT ")
        // more tables

Migrations that have to be performed:

    .upgradeTo(7, recreate(Tables.STATS))

And downgrades that you need to apply to get the previous db schema:

        new TableDowngrade(Tables.ADDRESSES,
            new DropColumn(Addresses.REGION)
    .downgradeTo(2, dropTable(Tables.ADDRESSES))

This might look like a tedious, unnecessary work. In reality it is tedious, but very helpful work. It reduces the usual db.execSQL() boilerplate in onCreate and onUpgrade to this:

public void onCreate(SQLiteDatabase db) {
  Schema currentSchema = SCHEMA.getCurrentSchema();
  for (String table : currentSchema.getTables()) {

public void onUpgrade(final SQLiteDatabase db, int oldVersion, int newVersion) {
  SCHEMA.upgrade(oldVersion, mContext, db);

Bulletproof workaround for SQLite's ALTER TABLE deficiencies

SQLite's ALTER TABLE supports only renaming the table or adding a new column. If you want to do anything more fancy like dropping the column, you have to do the following things:

  • Rename table with old schema to some temporary name
  • Create table with new schema
  • Move the data from renamed table to new table
  • Drop the old table

This library provides tested wrapper for this boilerplate with nice API:

TableMigration migration = TableMigration
    .withMapping(MyTable.SOME_COLUMN, "NULL")
migrationHelper.performMigrations(db, migration);

Warning: the MigrationHelper is not thread-safe (but seriously, why on earth would you want to perform sequential schema migrations in parallel?).

Write only non-trivial migrations

Most of the migrations you perform are trivial: dropping column, adding nullable column, adding table, dropping table, etc. If you specify complete schema history (which you should do anyways), this library will figure out trivial migrations for you. Of course you can still define fully custom migrations or just combine your custom migrations with our automagic:

    new SimpleMigration() {
      public void apply(SQLiteDatabase db, Schema schema) {
        // usual db.execSQL() 

Reduce merge conflicts

In your Schemas definition you can include release checkpoints. All revision numbers before this checkpoint are in fact offsets from this revision. It helps a lot when you are merging two branches, which introduced changes to your schema.

You still have to merge the section of code with currentSchema and you still have to make sure that both branches haven't performed the same changes, but you don't have to juggle the revision numbers and in 90% of cases you just need to decide which batch of changes should go first.

Automatic db index creation

Define the relationships between your data models using Thneed and use this information to generate the proper indexes:

for (SQLiteIndex index : AutoIndexer.generateIndexes(MODEL_GRAPH)) {

Note that this will generate the indexes for both ends of the relationships, which might not be exactly what you want. For example it will generate the index for primary keys referenced from other columns. We provide the Predicate factory to filter the generation results:

FluentIterable<SQLiteIndex> indexes = FluentIterable

And if you don't want index generation, you can still use our API to get the create index statement (although, I admit, it's not a killer feature):

AutoIndexer.getCreateStatement(new SQLiteIndex("my_table", "foobar_id"));

Hint: when you're automagically generating the indexes, you want to automagically clean them up as well. Use SQLiteMaster to do this:



Set of utils for getting existing db schema information from sqlite_master table.

You can use the schema information in your SQLiteOpenHelper's onCreate and onUpgrade to remove some boilerplate code. Compare:

db.execSQL("DROP TRIGGER IF EXISTS trigger_a");
db.execSQL("DROP TRIGGER IF EXISTS trigger_b");
db.execSQL("DROP TRIGGER IF EXISTS trigger_c");
// ...
db.execSQL("DROP TRIGGER IF EXISTS trigger_z");



You can perform similar operations with views, tables and indexes, or you can access the full schema information using getSQLiteSchemaParts(SQLiteDatabase db, SQLiteSchemaPartType partType) or getSQLiteSchemaParts(SQLiteDatabase db), which return the list of SQLiteSchemaPart objects:

public class SQLiteSchemaPart {
  public final String name;
  public final String sql;
  public final String type;

What you do with that information is completely up to you.

Is it safe to use?

Our tests indicate that there are no issues whatsoever on API level 8+ (Android 2.2). We haven't tested earlier versions, so consider yourself warned (and please let us know if you confirm it works on lower API levels!).


Just add the dependency to your build.gradle:

dependencies {
    compile ''

Copyright and license

Copyright 2013 Zendesk

Licensed under the Apache License, Version 2.0