SQLite/Ruby FAQ

How do I do a database query? I just want an array of the rows…

Use the Database#execute method. If you don’t give it a block, it will return an array of all the rows:

  require 'sqlite'

  db = SQLite::Database.new( "test.db" )
  rows = db.execute( "select * from test" )
How do I do a database query? I’d like to use a block to iterate through the rows…

Use the Database#execute method. If you give it a block, each row of the result will be yielded to the block:

  require 'sqlite'

  db = SQLite::Database.new( "test.db" )
  db.execute( "select * from test" ) do |row|
    ...
  end
How do I do a database query? I need to get the column names as well as the rows…

Use the Database#execute2 method. This works just like Database#execute; if you don’t give it a block, it returns an array of rows; otherwise, it will yield each row to the block. However, the first row returned is always an array of the column names from the query:

  require 'sqlite'

  db = SQLite::Database.new( "test.db" )
  columns, *rows = db.execute2( "select * from test" )

  # or use a block:

  columns = nil
  db.execute2( "select * from test" ) do |row|
    if columns.nil?
      columns = row
    else
      # process row
    end
  end
How do I do a database query? I need the result set object itself…

Sometimes you don’t want all the rows at once, and yet you’d like to be able to iterate through the results. For instance, you may want to pass the results to some other function (or series of functions) and have them pull rows from the results on demand. This is more effecient for very large queries.

To do this, use Database#query. If called without a block, it returns the ResultSet instance. If called with a block, it yields the ResultSet to the block (and then closes the set when the block terminates).

You can do both internal and external iteration with the result set this way:

  require 'sqlite'

  db = SQLite::Database.new( "test.db" )
  result = db.query( "select * from test" )

  if do_external_iteration

    while ( row = result.next )
      # process row
    end

  elsif do_internal_iteration

    result.each do |row|
      # process row
    end

  elsif get_result_metadata

    column_names = result.columns
    column_types = result.types

  end

  result.close
Note: if you are using Database#query to execute statements with no result sets (ie, inserts, deletes, udpates, etc.), you must call result.next to actually perform the operation. Otherwise, the operation will never be executed:
  db.query( "insert into table values ( 'a', 'b' )" ) do |result|
    result.next
  end

In general, Database#query is not a very good choice for such operations…

How do I do a database query? I just want the first row of the result set…

Easy. Just call Database#get_first_row:

  row = db.get_first_row( "select * from table" )

This also supports bind variables, just like Database#execute and friends.

How do I do a database query? I just want the first value of the first row of the result set…

Also easy. Just call Database#get_first_value:

  count = db.get_first_value( "select count(*) from table" )

This also supports bind variables, just like Database#execute and friends.

How do I prepare a statement for repeated execution?

If the same statement is going to be executed repeatedly, you can speed things up a bit by preparing the statement. You do this via the Database#prepare method. It returns a Statement object, and you can then invoke #execute on that to get the ResultSet:

  stmt = db.prepare( "select * from person" )

  1000.times do
    stmt.execute do |result|
      ...
    end
  end

This is made more useful by the ability to bind variables to placeholders via the Statement#bind_param and Statement#bind_params methods. (See the next FAQ for details.)

How do I use placeholders in an SQL statement?

Placeholders in an SQL statement take any of the following formats:

Where n is an integer, and word is an alpha-numeric identifier (or number). When the placeholder is associated with a number, that number identifies the index of the bind variable to replace it with. When it is an identifier, it identifies the name of the correponding bind variable. (In the instance of the first format—a single question mark—the placeholder is assigned a number one greater than the last index used, or 1 if it is the first.)

For example, here is a query using these placeholder formats:

  select *
    from table
   where ( c = ?2 or c = ? )
     and d = :name
     and e = :spouse:
     and f = :1

This defines 5 different placeholders: 1, 2, 3, “name”, and “spouse”.

You replace these placeholders by binding them to values. This can be accomplished in a variety of ways.

The Database#execute, Database#execute2, and Database#query methods all accept additional arguments following the SQL statement. These arguments are assumed to be bind parameters, and they are bound (positionally) to their corresponding placeholders:

  db.execute( "select * from table where a = ? and b = ?",
              "hello",
              "world" )

The above would replace the first question mark with ‘hello’ and the second with ‘world’. If the placeholders have an explicit index given, they will be replaced with the bind parameter at that index (1-based).

If a Hash is given as a bind parameter, then its key/value pairs are bound to the placeholders. This is how you bind by name:

  db.execute( "select * from table where a = :name and b = :value",
              "name" => "bob",
              "value" => "priceless" )

You can also bind explicitly using the Statement object itself. Just do a Database#prepare to get the Statement, and then use either Statement#bind_param or Statement#bind_params:

  stmt = db.prepare( "select * from table where a = :name and b = ?" )

  stmt.bind_param( "name", "bob" )
  stmt.bind_param( 1, "value" )

  # or

  stmt.bind_params( "value", "name" => "bob" )
How do I discover metadata about a query?

If you ever want to know the names or types of the columns in a result set, you can do it in several ways.

The first way is to ask the row object itself. Each row will have a property “fields” that returns an array of the column names. The row will also have a property “types” that returns an array of the column types:

  rows = db.execute( "select * from table" )
  p rows[0].fields
  p rows[0].types

Obviously, this approach requires you to execute a statement that actually returns data. If you don’t know if the statement will return any rows, but you still need the metadata, you can use Database#query and ask the ResultSet object itself:

  db.query( "select * from table" ) do |result|
    p result.columns
    p result.types
    ...
  end

Lastly, you can use Database#prepare and ask the Statement object what the metadata are:

  stmt = db.prepare( "select * from table" )
  p stmt.columns
  p stmt.types
I’d like the rows to be indexible by column name.

By default, each row from a query is returned as an Array of values. This means that you can only obtain values by their index. Sometimes, however, you would like to obtain values by their column name.

The first way to do this is to set the Database property “results_as_hash” to true. If you do this, then all rows will be returned as Hash objects, with the column names as the keys. (In this case, the “fields” property is unavailable on the row, although the “types” property remains.)

  db.results_as_hash = true
  db.execute( "select * from table" ) do |row|
    p row['column1']
    p row['column2']
  end

The other way is to use Ara Howard’s ArrayFields module. Just require “arrayfields”, and all of your rows will be indexable by column name, even though they are still arrays!

  require 'arrayfields'

  ...
  db.execute( "select * from table" ) do |row|
    p row[0] == row['column1']
    p row[1] == row['column2']
  end
I’d like the values from a query to be the correct types, instead of String.

You can turn on “type translation” by setting Database#type_translation to true:

  db.type_translation = true
  db.execute( "select * from table" ) do |row|
    p row
  end

By doing this, each return value for each row will be translated to its correct type, based on its declared column type.

You can even declare your own translation routines, if (for example) you are using an SQL type that is not handled by default:

  # assume "objects" table has the following schema:
  #   create table objects (
  #     name varchar2(20),
  #     thing object
  #   )

  db.type_translation = true
  db.translator.add_translator( "object" ) do |type, value|
    db.decode( value )
  end

  h = { :one=>:two, "three"=>"four", 5=>6 }
  dump = db.encode( h )

  db.execute( "insert into objects values ( ?, ? )", "bob", dump )

  obj = db.get_first_value( "select thing from objects where name='bob'" )
  p obj == h
How do I do a DDL (insert, update, delete) statement?

You can actually do inserts, updates, and deletes in exactly the same way as selects, but in general the Database#execute method will be most convenient:

  db.execute( "insert into table values ( ?, ? )", *bind_vars )
How do I execute multiple statements in a single string?

The standard query methods (Database#execute, Database#execute2, Database#query, and Statement#execute) will only execute the first statement in the string that is given to them. Thus, if you have a string with multiple SQL statements, each separated by a string, you can’t use those methods to execute them all at once.

Instead, use Database#execute_batch:

  sql = <<SQL
    create table the_table (
      a varchar2(30),
      b varchar2(30)
    );

    insert into the_table values ( 'one', 'two' );
    insert into the_table values ( 'three', 'four' );
    insert into the_table values ( 'five', 'six' );
  SQL

  db.execute_batch( sql )

Unlike the other query methods, Database#execute_batch accepts no block. It will also only ever return nil. Thus, it is really only suitable for batch processing of DDL statements.

How do I begin/end a transaction?

Use Database#transaction to start a transaction. If you give it a block, the block will be automatically committed at the end of the block, unless an exception was raised, in which case the transaction will be rolled back. (Never explicitly call Database#commit or Database#rollback inside of a transaction block—you’ll get errors when the block terminates!)

 database.transaction do |db| db.execute( "insert into table values ( 'a', 'b', 'c' )" ) ... end 

Alternatively, if you don’t give a block to Database#transaction, the transaction remains open until you explicitly call Database#commit or Database#rollback.

 db.transaction db.execute( "insert into table values ( 'a', 'b', 'c' )" ) db.commit 

Note that SQLite does not allow nested transactions, so you’ll get errors if you try to open a new transaction while one is already active. Use Database#transaction_active? to determine whether a transaction is active or not.