You are browsing a version that is no longer maintained. |
Data Retrieval And Manipulation
Doctrine DBAL follows the PDO API very closely. If you have worked with PDO before you will get to know Doctrine DBAL very quickly. On top of the API provided by PDO there are tons of convenience functions in Doctrine DBAL.
Data Retrieval
Using a database implies retrieval of data. It is the primary use-case of a database. For this purpose each database vendor exposes a Client API that can be integrated into programming languages. PHP has a generic abstraction layer for this kind of API called PDO (PHP Data Objects). However because of disagreements between the PHP community there are often native extensions for each database vendor that are much more maintained (OCI8 for example).
Doctrine DBAL API builds on top of PDO and integrates native extensions by
wrapping them into the PDO API as well. If you already have an open connection
through the Doctrine\DBAL\DriverManager::getConnection()
method you
can start using this API for data retrieval easily.
Start writing an SQL query and pass it to the query()
method of your
connection:
The query method executes the SQL and returns a database statement object. A database statement object can be iterated to retrieve all the rows that matched the query until there are no more rows:
The query method is the most simple one for fetching data, but it also has several drawbacks:
- There is no way to add dynamic parameters to the SQL query without modifying
$sql
itself. This can easily lead to a category of security holes called SQL injection, where a third party can modify the SQL executed and even execute their own queries through clever exploiting of the security hole. - Quoting dynamic parameters for an SQL query is tedious work and requires lots
of use of the
Doctrine\DBAL\Connection#quote()
method, which makes the original SQL query hard to read/understand. - Databases optimize SQL queries before they are executed. Using the query method you will trigger the optimization process over and over again, although it could re-use this information easily using a technique called prepared statements.
These three arguments and some more technical details hopefully convinced you to investigate prepared statements for accessing your database.
Dynamic Parameters and Prepared Statements
Consider the previous query, now parameterized to fetch only a single article by id.
Using ext/mysql (still the primary choice of MySQL access for many developers) you had to escape
every value passed into the query using mysql_real_escape_string()
to avoid SQL injection:
If you start adding more and more parameters to a query (for example in UPDATE or INSERT statements) this approach might lead to complex to maintain SQL queries. The reason is simple, the actual SQL query is not clearly separated from the input parameters. Prepared statements separate these two concepts by requiring the developer to add placeholders to the SQL query (prepare) which are then replaced by their actual values in a second step (execute).
Placeholders in prepared statements are either simple positional question marks (?
) or named labels starting with
a colon (e.g. :name1
). You cannot mix the positional and the named approach. You have to bind a parameter
to each placeholder.
The approach using question marks is called positional, because the values are bound in order from left to right
to any question mark found in the previously prepared SQL query. That is why you specify the
position of the variable to bind into the bindValue()
method:
Named parameters have the advantage that their labels can be re-used and only need to be bound once:
The following section describes the API of Doctrine DBAL with regard to prepared statements.
Support for positional and named prepared statements varies between the different database extensions. PDO implements its own client side parser so that both approaches are feasible for all PDO drivers. OCI8/Oracle only supports named parameters, but Doctrine implements a client side parser to allow positional parameters also. |
Using Prepared Statements
There are three low-level methods on Doctrine\DBAL\Connection
that allow you to
use prepared statements:
prepare($sql)
- Create a prepared statement of the typeDoctrine\DBAL\Statement
. Using this method is preferred if you want to re-use the statement to execute several queries with the same SQL statement only with different parameters.executeQuery($sql, $params, $types)
- Create a prepared statement for the passed SQL query, bind the given params with their binding types and execute the query. This method returns the executed prepared statement for iteration and is useful for SELECT statements.executeStatement($sql, $params, $types)
- Create a prepared statement for the passed SQL query, bind the given params with their binding types and execute the query. This method returns the number of affected rows by the executed query and is useful for UPDATE, DELETE and INSERT statements.
A simple usage of prepare was shown in the previous section, however it is useful to
dig into the features of a Doctrine\DBAL\Statement
a little bit more. There are essentially
two different types of methods available on a statement. Methods for binding parameters and types
and methods to retrieve data from a statement.
bindValue($pos, $value, $type)
- Bind a given value to the positional or named parameter in the prepared statement.bindParam($pos, &$param, $type)
- Bind a given reference to the positional or named parameter in the prepared statement.
If you are finished with binding parameters you have to call execute()
on the statement, which
will trigger a query to the database. After the query is finished you can access the results
of this query using the fetch API of a statement:
fetch($fetchStyle)
- Retrieves the next row from the statement or false if there are none. Moves the pointer forward one row, so that consecutive calls will always return the next row.fetchColumn($column)
- Retrieves only one column of the next row specified by column index. Moves the pointer forward one row, so that consecutive calls will always return the next row.fetchAll($fetchStyle)
- Retrieves all rows from the statement.
The fetch API of a prepared statement obviously works only for SELECT
queries.
If you find it tedious to write all the prepared statement code you can alternatively use
the Doctrine\DBAL\Connection#executeQuery()
and Doctrine\DBAL\Connection#executeStatement()
methods. See the API section below on details how to use them.
Additionally there are lots of convenience methods for data-retrieval and manipulation on the Connection, which are all described in the API section below.
Binding Types
Doctrine DBAL extends PDOs handling of binding types in prepared statements
considerably. Besides Doctrine\DBAL\ParameterType
constants, you
can make use of two very powerful additional features.
DoctrineDBALTypes Conversion
If you don't specify an integer (through one of Doctrine\DBAL\ParameterType
constants) to
any of the parameter binding methods but a string, Doctrine DBAL will
ask the type abstraction layer to convert the passed value from
its PHP to a database representation. This way you can pass \DateTime
instances to a prepared statement and have Doctrine convert them
to the appropriate vendors database format:
If you take a look at Doctrine\DBAL\Types\DateTimeType
you will see that
parts of the conversion are delegated to a method on the current database platform,
which means this code works independent of the database you are using.
Be aware this type conversion only works with |
List of Parameters Conversion
This is a Doctrine 2.1 feature. |
One rather annoying bit of missing functionality in SQL is the support for lists of parameters. You cannot bind an array of values into a single prepared statement parameter. Consider the following very common SQL statement:
1 SELECT * FROM articles WHERE id IN (?)
Since you are using an IN
expression you would really like to use it in the following way
(and I guess everybody has tried to do this once in his life, before realizing it doesn't work):
Implementing a generic way to handle this kind of query is tedious work. This is why most developers fallback to inserting the parameters directly into the query, which can open SQL injection possibilities if not handled carefully.
Doctrine DBAL implements a very powerful parsing process that will make this kind of prepared statement possible natively in the binding type system. The parsing necessarily comes with a performance overhead, but only if you really use a list of parameters. There are two special binding types that describe a list of integers or strings:
\Doctrine\DBAL\Connection::PARAM_INT_ARRAY
\Doctrine\DBAL\Connection::PARAM_STR_ARRAY
Using one of these constants as a type you can activate the SQLParser inside Doctrine that rewrites the SQL and flattens the specified values into the set of parameters. Consider our previous example:
The SQL statement passed to Connection#executeQuery
is not the one actually passed to the
database. It is internally rewritten to look like the following explicit code that could
be specified as well:
1 <?php
// Same SQL WITHOUT usage of Doctrine\DBAL\Connection::PARAM_INT_ARRAY
$stmt = $conn->executeQuery('SELECT * FROM articles WHERE id IN (?, ?, ?, ?, ?, ?)',
array(1, 2, 3, 4, 5, 6),
array(
ParameterType::INTEGER,
ParameterType::INTEGER,
ParameterType::INTEGER,
ParameterType::INTEGER,
ParameterType::INTEGER,
ParameterType::INTEGER,
)
);
2
3
4
5
6
7
8
9
10
11
12
13
This is much more complicated and is ugly to write generically.
The parameter list support only works with |
API
The DBAL contains several methods for executing queries against your configured database for data retrieval and manipulation. Below we'll introduce these methods and provide some examples for each of them.
executeStatement()
Executes a prepared statement with the given SQL and parameters and returns the affected rows count:
The $types
variable contains the PDO or Doctrine Type constants
to perform necessary type conversions between actual input
parameters and expected database values. See the
Types section for more information.
executeQuery()
Creates a prepared statement for the given SQL and passes the parameters to the execute method, then returning the statement:
The $types
variable contains the PDO or Doctrine Type constants
to perform necessary type conversions between actual input
parameters and expected database values. See the
Types section for more information.
fetchAllKeyValue()
Execute the query and fetch the first two columns into an associative array as keys and values respectively:
All additional columns will be ignored and are only allowed to be selected by DBAL for its internal purposes. |
fetchAllAssociativeIndexed()
Execute the query and fetch the data as an associative array where the key represents the first column and the value is an associative array of the rest of the columns and their values:
fetchAssociative()
Retrieve associative array of the first result row.
There are also convenience methods for data manipulation queries:
iterateKeyValue()
Execute the query and iterate over the first two columns as keys and values respectively:
All additional columns will be ignored and are only allowed to be selected by DBAL for its internal purposes. |
iterateAssociativeIndexed()
Execute the query and iterate over the result with the key representing the first column and the value being an associative array of the rest of the columns and their values:
update()
Update all rows for the matching key value identifiers with the given data.
By default the Doctrine DBAL does no escaping. Escaping is a very tricky business to do automatically, therefore there is none by default. The ORM internally escapes all your values, because it has lots of metadata available about the current context. When you use the Doctrine DBAL as standalone, you have to take care of this yourself. The following methods help you with it: