PyQtを使用したSQLデータベースの処理：基本

The final step is to bind the values that you want to pass to each placeholder and then call .exec() to execute the query. To do that, you use a for loop. The loop header unpacks each tuple in data into three separate variables with convenient names. Then you call .addBindValue() on the query object to bind the values to the placeholders.

Note that you’re using positional placeholders , so the order in which you call .addBindValue() will define the order in which each value is passed to the corresponding placeholder.

This approach for creating dynamic queries is handy when you want to customize your queries using values that come from your user’s input. Anytime you take the user’s input to complete a query on a database, you face the security risk of SQL injection.

In PyQt, combining .prepare() , .bindValue() , and .addBindValue() fully protects you against SQL injection attacks, so this is the way to go when you’re taking untrusted input to complete your queries.

Navigating the Records in a Query

If you execute a SELECT statement, then your QSqlQuery object will retrieve zero or more records from one or more tables in your database. The query will hold records containing data that matches the query’s criteria. If no data matches the criteria, then your query will be empty.

QSqlQuery provides a set of navigation methods that you can use to move throughout the records in a query result:

All these methods position the query object on the retrieved record if that record is available. Most of these methods have specific rules that apply when using them. With these methods, you can move forward, backward, or arbitrarily through the records in a query result. Since they all return either True or False , you can use them in a while loop to navigate all the records in one go.

These methods work with active queries 。 A query is active when you’ve successfully run .exec() on it, but the query isn’t finished yet. Once an active query is on a valid record, you can retrieve data from that record using .value(index) 。 This method takes a zero-based integer number, index , and returns the value at that index (column) in the current record.

Note: If you execute a SELECT * type of query, then the columns in the result won’t follow a known order. This might cause problems when you use .value() to retrieve the value at a given column because there’s no way of knowing if you’re using the right column index.

You’ll look at a few examples of how to use some of the navigation methods to move throughout a query below. But first, you need to create a connection to your database:

>>> from PyQt5.QtSql import QSqlDatabase, QSqlQuery

>>> con = QSqlDatabase.addDatabase("QSQLITE")
>>> con.setDatabaseName("contacts.sqlite")
>>> con.open()
True

Here, you create and open a new connection to contacts.sqlite 。 If you’ve been following along with this tutorial so far, then this database already contains some sample data. Now you can create a QSqlQuery object and execute it on that data:

>>> # Create and execute a query
>>> query = QSqlQuery()
>>> query.exec("SELECT name, job, email FROM contacts")
True

This query retrieves data about the name , job , and email of all the contacts stored in the contacts table. Since .exec() returned True , the query was successful and is now an active query. You can navigate the records in this query using any of the navigation methods you saw before. You can also retrieve the data at any column in a record using .value() ：

>>> # First record
>>> query.first()
True

>>> # Named indices for readability
>>> name, job, email = range(3)

>>> # Retrieve data from the first record
>>> query.value(name)
'Linda'

>>> # Next record
>>> query.next()
True
>>> query.value(job)
'Senior Web Developer'

>>> # Last record
>>> query.last()
True
>>> query.value(email)
'[email protected]'

With the navigation methods, you can move around the query result. With .value() , you can retrieve the data at any column in a given record.

You can also iterate through all the records in your query using a while loop along with .next() ：

>>> query.exec()
True

>>> while query.next():
...     print(query.value(name), query.value(job), query.value(email))
...
Linda Technical Lead [email protected]
Joe Senior Web Developer [email protected]
...

With .next() , you navigate all the records in a query result. .next() works similar to the iterator protocol in Python. Once you’ve iterated over the records in a query result, .next() starts returning False until you run .exec() また。 A call to .exec() retrieves data from a database and places the query object’s internal pointer one position before the first record, so when you call .next() , you get the first record again.

You can also loop in reverse order using .previous() ：

>>> while query.previous():
...     print(query.value(name), query.value(job), query.value(email))
...
Jane Senior Python Developer [email protected]
David Data Analyst [email protected]
...

.previous() works similar to .next() , but the iteration is done in reverse order. In other words, the loop goes from the query pointer’s position back to the first record.

Sometimes you might want to get the index that identifies a given column in a table by using the name of that column. To do that, you can call .indexOf() on the return value of .record() ：

>>> query.first()
True

>>> # Get the index of name
>>> name = query.record().indexOf("name")

>>> query.value(name)
'Linda'

>>> # Finish the query object if unneeded
>>> query.finish()
>>> query.isActive()
False

The call to .indexOf() on the result of .record() returns the index of the "name" 桁。 If "name" doesn’t exist, then .indexOf() returns -1 。 This is handy when you use a SELECT * statement in which the order of columns is unknown. Finally, if you’re done with a query object, then you can turn it inactive by calling .finish() 。 This will free the system memory associated with the query object at hand.

Closing and Removing Database Connections

In practice, some of your PyQt applications will depend on a database, and others won’t. An application that depends on a database often creates and opens a database connection just before creating any window or graphical component and keeps the connection open until the application is closed.

On the other hand, applications that don’t depend on a database but use a database to provide some of their functionalities typically connect to that database only when needed, if at all. In these cases, you can close the connection after use and free the resources associated with that connection, such as system memory.

To close a connection in PyQt, you call .close() on the connection. This method closes the connection and frees any acquired resources. It also invalidates any associated QSqlQuery objects because they can’t work properly without an active connection.

Here’s an example of how to close an active database connection using .close() ：

>>> from PyQt5.QtSql import QSqlDatabase

>>> con = QSqlDatabase.addDatabase("QSQLITE")
>>> con.setDatabaseName("contacts.sqlite")

>>> con.open()
True
>>> con.isOpen()
True

>>> con.close()
>>> con.isOpen()
False

You can call .close() on a connection to close it and free all its associated resources. To make sure that a connection is closed, you call .isOpen() 。

Note that QSqlQuery objects remain in memory after closing their associated connection, so you must make your queries inactive by calling .finish() or .clear() , or by deleting the QSqlQuery object before closing the connection. Otherwise, residual memory is left out in your query object.

You can reopen and reuse any previously closed connection. That’s because .close() doesn’t remove connections from the list of available connections, so they remain usable.

You can also completely remove your database connections using .removeDatabase() 。 To do this safely, first finish your queries using .finish() , then close the database using .close() , and finally remove the connection. You can use .removeDatabase(connectionName) to remove the database connection called connectionName from the list of available connections. Removed connections are no longer available for use in the application at hand.

To remove the default database connection, you can call .connectionName() on the object returned by .database() and pass the result to .removeDatabase() ：

>>> # The connection is closed but still in the list of connections
>>> QSqlDatabase.connectionNames()
['qt_sql_default_connection']

>>> # Remove the default connection
>>> QSqlDatabase.removeDatabase(QSqlDatabase.database().connectionName())

>>> # The connection is no longer in the list of connections
>>> QSqlDatabase.connectionNames()
[]

>>> # Try to open a removed connection
>>> con.open()
False

Here, the call to .connectionNames() returns the list of available connections. In this case, you have only one connection, the default. Then you remove the connection using .removeDatabase() 。

Note: Before closing and removing a database connection, you need to make sure that everything that uses the connection is deleted or set to use a different data source. Otherwise, you can have a resource leak 。

Since you need a connection name to use .removeDatabase() , you call .connectionName() on the result of .database() to get the name of the default connection. Finally, you call .connectionNames() again to make sure that the connection is no longer in the list of available connections. Trying to open a removed connection will return False because the connection no longer exists.

Displaying and Editing Data With PyQt

A common requirement in GUI applications that use databases is the ability to load, display, and edit data from the database using different widgets. Table, list, and tree widgets are commonly used in GUIs to manage data.

PyQt provides two different kind of widgets for managing data:

1. Standard widgets include internal containers for storing data.
2. View widgets don’t maintain internal data containers but use models to access data.

For small GUI applications that manage small databases, you can use the first approach. The second approach is handy when you’re building complex GUI applications that manage large databases.

The second approach takes advantage of PyQt’s Model-View programming. With this approach, you have widgets that represent views such as tables, lists, and trees on one hand and model classes that communicate with your data on the other hand.

Understanding PyQt’s Model-View Architecture

The Model-View-Controller (MVC) design pattern is a general software pattern intended to divide an application’s code into three general layers, each with a different role.

The model takes care of the business logic of the application, the view provides on-screen representations, and the controller connects the model and the view to make the application work.

Qt provides a custom variation of MVC. They call it the Model-View architecture, and it’s available for PyQt as well. The pattern also separates the logic into three components:

1. Models communicate with and access the data. They also define an interface that’s used by views and delegates to access the data. All models are based on QAbstractItemModel 。 Some commonly used models include QStandardItemModel , QFileSystemModel , and SQL-related models.

2. Views are responsible for displaying the data to the user. They also have similar functionality to the controller in the MVC pattern. All views are based on QAbstractItemView 。 Some commonly used views are QListView , QTableView , and QTreeView 。

3. Delegates paint view items and provide editor widgets for modifying items. They also communicate back with the model if an item has been modified. The base class is QAbstractItemDelegate 。

Separating classes into these three components implies that changes on models will be reflected on associated views or widgets automatically, and changes on views or widgets through delegates will update the underlying model automatically.

In addition, you can display the same data in different views without the need for multiple models.

Using Standard Widget Classes

PyQt provides a bunch of standard widgets for displaying and editing data in your GUI applications. These standard widgets provide views such as tables, trees, and lists. They also provide an internal container for storing data and convenient delegates for editing the data. All these features are grouped into a single class.

Here are three of these standard classes:

QTableWidget is arguably the most popular widget when it comes to displaying and editing data. It creates a 2D array of QTableWidgetItem objects. Each item holds an individual value as a string. All these values are displayed and organized in a table of rows and columns.

You can perform at least the following operations on a QTableWidget object:

• Editing the content of its items using delegate objects
• Adding new items using .setItem()
• Setting the number of rows and columns using .setRowCount() and .setColumnCount()
• Adding vertical and horizontal header labels using setHorizontalHeaderLabels() and .setVerticalHeaderLabels

Here’s a sample application that shows how to use a QTableWidget object to display data in a GUI. The application uses the database you created and populated in previous sections, so if you want to run it, then you need to save the code into the same directory in which you have the contacts.sqlite database:

If you double-click any cell of the table, then you’ll be able to edit the content of the cell. However, your changes won’t be saved to your database.

Here’s the code for your application:

 1import sys
 2
 3from PyQt5.QtSql import QSqlDatabase, QSqlQuery
 4from PyQt5.QtWidgets import (
 5    QApplication,
 6    QMainWindow,
 7    QMessageBox,
 8    QTableWidget,
 9    QTableWidgetItem,
10)
11
12class Contacts(QMainWindow):
13    def __init__(self, parent=None):
14        super().__init__(parent)
15        self.setWindowTitle("QTableView Example")
16        self.resize(450, 250)
17        # Set up the view and load the data
18        self.view = QTableWidget()
19        self.view.setColumnCount(4)
20        self.view.setHorizontalHeaderLabels(["ID", "Name", "Job", "Email"])
21        query = QSqlQuery("SELECT id, name, job, email FROM contacts")
22        while query.next():
23            rows = self.view.rowCount()
24            self.view.setRowCount(rows + 1)
25            self.view.setItem(rows, 0, QTableWidgetItem(str(query.value(0))))
26            self.view.setItem(rows, 1, QTableWidgetItem(query.value(1)))
27            self.view.setItem(rows, 2, QTableWidgetItem(query.value(2)))
28            self.view.setItem(rows, 3, QTableWidgetItem(query.value(3)))
29        self.view.resizeColumnsToContents()
30        self.setCentralWidget(self.view)
31
32def createConnection():
33    con = QSqlDatabase.addDatabase("QSQLITE")
34    con.setDatabaseName("contacts.sqlite")
35    if not con.open():
36        QMessageBox.critical(
37            None,
38            "QTableView Example - Error!",
39            "Database Error: %s" % con.lastError().databaseText(),
40        )
41        return False
42    return True
43
44app = QApplication(sys.argv)
45if not createConnection():
46    sys.exit(1)
47win = Contacts()
48win.show()
49sys.exit(app.exec_())

Here’s what’s happening in this example:

• Lines 18 to 20 create a QTableWidget object, set the number of columns to 4 , and set user-friendly labels for each column’s header.
• Line 21 creates and executes a SELECT SQL query on your database to get all the data in the contacts table.
• Line 22 starts a while loop to navigate the records in the query result using .next() 。
• Line 24 increments the number of rows in the table by 1 using .setRowCount() 。
• Lines 25 to 28 add items of data to your table using .setItem() 。 Note that since the values in the id columns are integer numbers, you need to convert them into strings to be able to store them in a QTableWidgetItem object.

.setItem() takes three arguments:

1. row holds a zero-based integer that represents the index of a given row in the table.
2. column holds a zero-based integer that represents the index of a given column in the table.
3. item holds the QTableWidgetItem object that you need to place at a given cell in the table.

Finally, you call .resizeColumnsToContents() on your view to adjust the size of the columns to their content and provide a better rendering of the data.

Displaying and editing database tables using standard widgets can become a challenging task. That’s because you’ll have two copies of the same data. In other words you’ll have a copy of the data in two locations:

1. Outside the widget, in your database
2. Inside the widget, in the widget’s internal containers

You’re responsible for synchronizing both copies of your data manually, which can be an annoying and error-prone operation. Luckily, you can use PyQt’s Model-View architecture to avoid most of these problems, as you’ll see in the following section.

Using View and Model Classes

PyQt’s Model-View classes eliminate the problems of data duplication and synchronization that may occur when you use standard widget classes to build database applications. The Model-View architecture allows you to use several views to display the same data because you can pass one model to many views.

Model classes provide an application programming interface (API) that you can use to manipulate data. View classes provide convenient delegate objects that you can use to edit data in the view directly. To connect a view with a given module, you need to call .setModel() on the view object.

PyQt offers a set of view classes that support the Model-View architecture:

You can use these view classes along with model classes to create your database applications. This will make your applications more robust, faster to code, and less error-prone.

Here are some of the model classes that PyQt provides for working with SQL databases:

Once you’ve connected one of these models to a physical database table or query, you can use them to populate your views. Views provide delegate objects that allow you to modify the data directly in the view. The model connected to the view will update the data in your database to reflect any change in the view. Note that you don’t have to update the data in the database manually. The model will do that for you.

Here’s an example that shows the basics of how to use a QTableView object and a QSqlTableModel object together to build a database application using PyQt’s Model-View architecture:

To edit the data in a cell of the table, you can double-click the cell. A convenient delegate widget will show in the cell, allowing you to edit its content. Then you can hit Enter to save the changes.

The ability to automatically handle and save changes in the data is one of the more important advantages of using PyQt’s Model-View classes. The Model-View architecture will improve your productivity and reduce the errors that can appear when you have to write data manipulation code by yourself.

Here’s the code to create the application:

 1import sys
 2
 3from PyQt5.QtCore import Qt
 4from PyQt5.QtSql import QSqlDatabase, QSqlTableModel
 5from PyQt5.QtWidgets import (
 6    QApplication,
 7    QMainWindow,
 8    QMessageBox,
 9    QTableView,
10)
11
12class Contacts(QMainWindow):
13    def __init__(self, parent=None):
14        super().__init__(parent)
15        self.setWindowTitle("QTableView Example")
16        self.resize(415, 200)
17        # Set up the model
18        self.model = QSqlTableModel(self)
19        self.model.setTable("contacts")
20        self.model.setEditStrategy(QSqlTableModel.OnFieldChange)
21        self.model.setHeaderData(0, Qt.Horizontal, "ID")
22        self.model.setHeaderData(1, Qt.Horizontal, "Name")
23        self.model.setHeaderData(2, Qt.Horizontal, "Job")
24        self.model.setHeaderData(3, Qt.Horizontal, "Email")
25        self.model.select()
26        # Set up the view
27        self.view = QTableView()
28        self.view.setModel(self.model)
29        self.view.resizeColumnsToContents()
30        self.setCentralWidget(self.view)
31
32def createConnection():
33    con = QSqlDatabase.addDatabase("QSQLITE")
34    con.setDatabaseName("contacts.sqlite")
35    if not con.open():
36        QMessageBox.critical(
37            None,
38            "QTableView Example - Error!",
39            "Database Error: %s" % con.lastError().databaseText(),
40        )
41        return False
42    return True
43
44app = QApplication(sys.argv)
45if not createConnection():
46    sys.exit(1)
47win = Contacts()
48win.show()
49sys.exit(app.exec_())

Here’s what’s happening in this code:

• Line 18 creates an editable QSqlTableModel object.
• Line 19 connects your model with the contacts table in your database using .setTable() 。
• Line 20 sets the edit strategy of the model to OnFieldChange 。 This strategy allows the model to automatically update the data in your database if the user modifies any of the data directly in the view.
• Lines 21 to 24 set some user-friendly labels to the horizontal headers of the model using .setHeaderData() 。
• Line 25 loads the data from your database and populates the model by calling .select() 。
• Line 27 creates the table view object to display the data contained in the model.
• Line 28 connects the view with the model by calling .setModel() on the view with your data model as an argument.
• Line 29 calls .resizeColumnsToContents() on the view object to adjust the table to its content.

That’s it! You now have a fully-functional database application.

Using SQL Databases in PyQt:Best Practices

When it comes to using PyQt’s SQL support effectively, there are some best practices that you might want to use in your applications:

• Favor PyQt’s SQL support over Python standard library or third-party libraries to take advantage of the natural integration of these classes with the rest of PyQt’s classes and infrastructure, mostly with the Model-View architecture.

• Use previously prepared dynamic queries with placeholders for parameters and bind values to those parameters using .addBindValue() and .bindValue() 。 This will help prevent SQL injection attacks.

• Handle errors that can occur when opening a database connection to avoid unexpected behaviors and application crashes.

• Close and remove unneeded database connections and queries to free any acquired system resources.

• Minimize the use of SELECT * queries to avoid problems when retrieving data with .value() 。

• Pass your passwords to .open() instead of to .setPassword() to avoid the risk of compromising your security.

• Take advantage of PyQt’s Model-View architecture and its integration with PyQt’s SQL support to make your applications more robust.

This list isn’t complete, but it’ll help you make better use of PyQt’s SQL support when developing your database applications.

Conclusion

Using PyQt’s built-in support to work with SQL databases is an important skill for any Python developer who’s creating PyQt GUI applications and needs to connect them to a database. PyQt provides a consistent set of classes for managing SQL databases.

These classes fully integrate with PyQt’s Model-View architecture, allowing you to develop GUI applications that can manage databases in a user-friendly way.

In this tutorial, you’ve learned how to:

• Use PyQt’s SQL support to connect to a database
• Execute SQL queries on a database with PyQt
• Build database applications using PyQt’s Model-View architecture
• Display and edit data from a database using PyQt widgets

With this knowledge, you can improve your productivity when creating nontrivial database applications and make your GUI applications more robust.