A TQSqlCursor is a database record (see TQSqlRecord) that corresponds to a table or view within an SQL database (see TQSqlDatabase). There are two buffers in a cursor, one used for browsing and one used for editing records. Each buffer contains a list of fields which correspond to the fields in the table or view.
When positioned on a valid record, the browse buffer contains the values of the current record's fields from the database. The edit buffer is separate, and is used for editing existing records and inserting new records.
.PP
For browsing data, a cursor must first select() data from the database. After a successful select() the cursor is active (isActive() returns TRUE), but is initially not positioned on a valid record (isValid() returns FALSE). To position the cursor on a valid record, use one of the navigation functions, next(), prev(), first(), last(), or seek(). Once positioned on a valid record, data can be retrieved from the browse buffer using value(). If a navigation function is not successful, it returns FALSE, the cursor will no longer be positioned on a valid record and the values returned by value() are undefined.
In the above example, a cursor is created specifying a table or view name in the database. Then, select() is called, which can be optionally parameterised to filter and order the records retrieved. Each record in the cursor is retrieved using next(). When next() returns FALSE, there are no more records to process, and the loop terminates.
For editing records (rows of data), a cursor contains a separate edit buffer which is independent of the fields used when browsing. The functions insert(), update() and del() operate on the edit buffer. This allows the cursor to be repositioned to other records while simultaneously maintaining a separate buffer for edits. You can get a pointer to the edit buffer using editBuffer(). The primeInsert(), primeUpdate() and primeDelete() functions also return a pointer to the edit buffer and prepare it for insert, update and delete respectively. Edit operations only affect a single row at a time. Note that update() and del() require that the table or view contain a primaryIndex() to ensure that edit operations affect a unique record within the database.
To edit an existing database record, first move to the record you wish to update. Call primeUpdate() to get the pointer to the cursor's edit buffer. Then use this pointer to modify the values in the edit buffer. Finally, call update() to save the changes to the database. The values in the edit buffer will be used to locate the appropriate record when updating the database (see primaryIndex()).
.PP
Similarly, when deleting an existing database record, first move to the record you wish to delete. Then, call primeDelete() to get the pointer to the edit buffer. Finally, call del() to delete the record from the database. Again, the values in the edit buffer will be used to locate and delete the appropriate record.
.PP
To insert a new record, call primeInsert() to get the pointer to the edit buffer. Use this pointer to populate the edit buffer with new values and then insert() the record into the database.
.PP
After calling insert(), update() or del(), the cursor is no longer positioned on a valid record and can no longer be navigated (isValid() return FALSE). The reason for this is that any changes made to the database will not be visible until select() is called to refresh the cursor. You can change this behavior by passing FALSE to insert(), update() or del() which will prevent the cursor from becoming invalid. The edits will still not be visible when navigating the cursor until select() is called.
TQSqlCursor contains virtual methods which allow editing behavior to be customized by subclasses. This allows custom cursors to be created that encapsulate the editing behavior of a database table for an entire application. For example, a cursor can be customized to always auto-number primary index fields, or provide fields with suitable default values, when inserting new records. TQSqlCursor generates SQL statements which are sent to the database engine; you can control which fields are included in these statements using setGenerated().
Note that TQSqlCursor does not inherit from TQObject. This means that you are responsible for destroying instances of this class yourself. However if you create a TQSqlCursor and use it in a TQDataTable, TQDataBrowser or a TQDataView these classes will usually take ownership of the cursor and destroy it when they don't need it anymore. The documentation for TQDataTable, TQDataBrowser and TQDataView explicitly states which calls take ownership of the cursor.
Constructs a cursor on database \fIdb\fR using table or view \fIname\fR.
.PP
If \fIautopopulate\fR is TRUE (the default), the \fIname\fR of the cursor must correspond to an existing table or view name in the database so that field information can be automatically created. If the table or view does not exist, the cursor will not be functional.
The cursor is created with an initial mode of TQSqlCursor::Writable (meaning that records can be inserted, updated or deleted using the cursor). If the cursor does not have a unique primary index, update and deletes cannot be performed.
Note that \fIautopopulate\fR refers to populating the cursor with meta-data, e.g. the names of the table's fields, not with retrieving data. The select() function is used to populate the cursor with data.
Protected virtual function which is called whenever a field needs to be calculated. If calculated fields are being used, derived classes must reimplement this function and return the appropriate value for field \fIname\fR. The default implementation returns an invalid TQVariant.
Deletes a record from the database using the cursor's primary index and the contents of the cursor edit buffer. Returns the number of records which were deleted. For error information, use lastError().
.PP
Only records which meet the filter criteria specified by the cursor's primary index are deleted. If the cursor does not contain a primary index, no delete is performed and 0 is returned. If \fIinvalidate\fR is TRUE (the default), the current cursor can no longer be navigated. A new select() call must be made before you can move to a valid record. For example:
In the above example, a cursor is created on the 'prices' table and positioned to the record to be deleted. First primeDelete() is called to populate the edit buffer with the current cursor values, e.g. with an id of 999, and then del() is called to actually delete the record from the database. Remember: all edit operations (insert(), update() and delete()) operate on the contents of the cursor edit buffer and not on the contents of the cursor itself.
.PP
See also primeDelete(), setMode(), and lastError().
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Deletes the current cursor record from the database using the filter \fIfilter\fR. Only records which meet the filter criteria are deleted. Returns the number of records which were deleted. If \fIinvalidate\fR is TRUE (the default), the current cursor can no longer be navigated. A new select() call must be made before you can move to a valid record. For error information, use lastError().
.PP
The \fIfilter\fR is an SQL \fCWHERE\fR clause, e.g. \fCid=500\fR.
Returns the current internal edit buffer. If \fIcopy\fR is TRUE (the default is FALSE), the current cursor field values are first copied into the edit buffer. The edit buffer is valid as long as the cursor remains valid. The cursor retains ownership of the returned pointer, so it must not be deleted or modified.
.PP
See also primeInsert(), primeUpdate(), and primeDelete().
Returns an index composed of \fIfieldNames\fR, all in ASCending order. Note that all field names must exist in the cursor, otherwise an empty index is returned.
Insert a copy of \fIfieldInfo\fR at position \fIpos\fR. If a field already exists at \fIpos\fR, it is removed. Note that all references to the cursor edit buffer become invalidated.
.PP
Examples:
.)l sql/overview/insert/main.cpp and sql/overview/insert2/main.cpp.
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Inserts the current contents of the cursor's edit record buffer into the database, if the cursor allows inserts. Returns the number of rows affected by the insert. For error information, use lastError().
.PP
If \fIinvalidate\fR is TRUE (the default), the cursor will no longer be positioned on a valid record and can no longer be navigated. A new select() call must be made before navigating to a valid record.
In the above example, a cursor is created on the 'prices' table and a pointer to the insert buffer is acquired using primeInsert(). Each field's value is set to the desired value and then insert() is called to insert the data into the database. Remember: all edit operations (insert(), update() and delete()) operate on the contents of the cursor edit buffer and not on the contents of the cursor itself.
Returns TRUE if the cursor is read-only; otherwise returns FALSE. The default is FALSE. Read-only cursors cannot be edited using insert(), update() or del().
Returns the primary index associated with the cursor as defined in the database, or an empty index if there is no primary index. If \fIsetFromCursor\fR is TRUE (the default), the index fields are populated with the corresponding values in the cursor's current record.
This function primes the edit buffer's field values for delete and returns the edit buffer. The default implementation copies the field values from the current cursor record into the edit buffer (therefore, this function is equivalent to calling editBuffer( TRUE ) ). The cursor retains ownership of the returned pointer, so it must not be deleted or modified.
This function primes the edit buffer's field values for insert and returns the edit buffer. The default implementation clears all field values in the edit buffer. The cursor retains ownership of the returned pointer, so it must not be deleted or modified.
.PP
See also editBuffer() and insert().
.PP
Examples:
.)l sql/overview/insert/main.cpp, sql/overview/insert2/main.cpp, sql/overview/subclass5/main.cpp, and sql/sqltable/main.cpp.
This function primes the edit buffer's field values for update and returns the edit buffer. The default implementation copies the field values from the current cursor record into the edit buffer (therefore, this function is equivalent to calling editBuffer( TRUE ) ). The cursor retains ownership of the returned pointer, so it must not be deleted or modified.
.PP
See also editBuffer() and update().
.PP
Examples:
.)l sql/overview/custom1/main.cpp, sql/overview/form1/main.cpp, and sql/overview/update/main.cpp.
Selects all fields in the cursor from the database matching the filter criteria \fIfilter\fR. The data is returned in the order specified by the index \fIsort\fR. Returns TRUE if the data was successfully selected; otherwise returns FALSE.
.PP
The \fIfilter\fR is a string containing a SQL \fCWHERE\fR clause but without the 'WHERE' keyword. The cursor is initially positioned at an invalid row after this function is called. To move to a valid row, use seek(), first(), last(), prev() or next().
cur.select( "deptno=10" ); // select all records in department 10
.br
while( cur.next() ) {
.br
... // process data
.br
}
.br
...
.br
// select records in other departments, ordered by department number
.br
cur.select( "deptno>10", cur.index( "deptno" ) );
.br
...
.br
.fi
.PP
The filter will apply to any subsequent select() calls that do not explicitly specify another filter. Similarly the sort will apply to any subsequent select() calls that do not explicitly specify another sort.
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Selects all fields in the cursor from the database. The rows are returned in the order specified by the last call to setSort() or the last call to select() that specified a sort, whichever is the most recent. If there is no current sort, the order in which the rows are returned is undefined. The records are filtered according to the filter specified by the last call to setFilter() or the last call to select() that specified a filter, whichever is the most recent. If there is no current filter, all records are returned. The cursor is initially positioned at an invalid row. To move to a valid row, use seek(), first(), last(), prev() or next().
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Selects all fields in the cursor from the database. The data is returned in the order specified by the index \fIsort\fR. The records are filtered according to the filter specified by the last call to setFilter() or the last call to select() that specified a filter, whichever is the most recent. The cursor is initially positioned at an invalid row. To move to a valid row, use seek(), first(), last(), prev() or next().
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Selects all fields in the cursor matching the filter index \fIfilter\fR. The data is returned in the order specified by the index \fIsort\fR. The \fIfilter\fR index works by constructing a WHERE clause using the names of the fields from the \fIfilter\fR and their values from the current cursor record. The cursor is initially positioned at an invalid row. To move to a valid row, use seek(), first(), last(), prev() or next(). This function is useful, for example, for retrieving data based upon a table's primary index:
In this example the TQSqlIndex, pk, is used for two different purposes. When used as the filter (first) argument, the field names it contains are used to construct the WHERE clause, each set to the current cursor value, \fCWHERE id=10\fR, in this case. When used as the sort (second) argument the field names it contains are used for the ORDER BY clause, \fCORDER BY id\fR in this example.
Sets field \fIname\fR to \fIcalculated\fR. If the field \fIname\fR does not exist, nothing happens. The value of a calculated field is set by the calculateField() virtual function which you must reimplement (or the field value will be an invalid TQVariant). Calculated fields do not appear in generated SQL statements sent to the database.
Sets the current filter to \fIfilter\fR. Note that no new records are selected. To select new records, use select(). The \fIfilter\fR will apply to any subsequent select() calls that do not explicitly specify a filter.
.PP
The filter is a SQL \fCWHERE\fR clause without the keyword 'WHERE', e.g. \fCname='Dave'\fR which will be processed by the DBMS.
Sets the generated flag for the field \fIname\fR to \fIgenerated\fR. If the field does not exist, nothing happens. Only fields that have \fIgenerated\fR set to TRUE are included in the SQL that is generated by insert(), update() or del().
.SH "void TQSqlCursor::setMode ( int mode )\fC [virtual]\fR"
Sets the cursor mode to \fImode\fR. This value can be an OR'ed combination of TQSqlCursor::Mode values. The default mode for a cursor is TQSqlCursor::Writable.
Sets the name of the cursor to \fIname\fR. If \fIautopopulate\fR is TRUE (the default), the \fIname\fR must correspond to a valid table or view name in the database. Also, note that all references to the cursor edit buffer become invalidated when fields are auto-populated. See the TQSqlCursor constructor documentation for more information.
Sets the primary index associated with the cursor to the index \fIidx\fR. Note that this index must contain a field or set of fields which identify a unique record within the underlying database table or view so that update() and del() will execute as expected.
Sets the current sort to \fIsort\fR. Note that no new records are selected. To select new records, use select(). The \fIsort\fR will apply to any subsequent select() calls that do not explicitly specify a sort.
Returns a formatted string composed of all the fields in \fIrec\fR. Each field is composed of the \fIprefix\fR (e.g. table or view name),"
.", the field name, the \fIfieldSep\fR and the field value. If the
\fIprefix\fR is empty then each field will begin with the field name. The fields are then joined together separated by \fIsep\fR. Fields where isGenerated() returns FALSE are not included. This function is useful for generating SQL statements.
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Returns a formatted string composed of the \fIprefix\fR (e.g. table or view name), ".", the \fIfield\fR name, the \fIfieldSep\fR and the field value. If the \fIprefix\fR is empty then the string will begin with the \fIfield\fR name. This function is useful for generating SQL statements.
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Returns a formatted string composed of all the fields in the index \fIi\fR. Each field is composed of the \fIprefix\fR (e.g. table or view name), ".", the field name, the \fIfieldSep\fR and the field value. If the \fIprefix\fR is empty then each field will begin with the field name. The field values are taken from \fIrec\fR. The fields are then joined together separated by \fIsep\fR. Fields where isGenerated() returns FALSE are ignored. This function is useful for generating SQL statements.
Updates the database with the current contents of the edit buffer. Returns the number of records which were updated. For error information, use lastError().
.PP
Only records which meet the filter criteria specified by the cursor's primary index are updated. If the cursor does not contain a primary index, no update is performed and 0 is returned.
.PP
If \fIinvalidate\fR is TRUE (the default), the current cursor can no longer be navigated. A new select() call must be made before you can move to a valid record. For example:
In the above example, a cursor is created on the 'prices' table and is positioned on the record to be updated. Then a pointer to the cursor's edit buffer is acquired using primeUpdate(). A new value is calculated and placed into the edit buffer with the setValue() call. Finally, an update() call is made on the cursor which uses the tables's primary index to update the record in the database with the contents of the cursor's edit buffer. Remember: all edit operations (insert(), update() and delete()) operate on the contents of the cursor edit buffer and not on the contents of the cursor itself.
This is an overloaded member function, provided for convenience. It behaves essentially like the above function.
.PP
Updates the database with the current contents of the cursor edit buffer using the specified \fIfilter\fR. Returns the number of records which were updated. For error information, use lastError().
.PP
Only records which meet the filter criteria are updated, otherwise all records in the table are updated.
.PP
If \fIinvalidate\fR is TRUE (the default), the cursor can no longer be navigated. A new select() call must be made before you can move to a valid record.
.PP
See also primeUpdate(), setMode(), and lastError().
