| Data Types | |
| enum StatementType |
#include <rw/db/osql.h>
Class RWDBOSql is an encapsulation of a database-specific SQL statement with all its input and output bindings. Program variables may be bound to placeholders in your statement using operator <<, and shifting in RWDBAbstractBuffers, such as RWDBTBuffers, which describe input data. Class RWDBOSql also contains a vector of RWDBMultiRows, each of which specifies the output bindings for a single result set. You can use the RWDBOSql operator[] to obtain a reference to an RWDBMultiRow, and then use the RWDBMultiRow operator>> to provide output bindings for the RWDBMultiRow.
When executing an RWDBOSql that produces results, there are two options for specifying the output RWDBAbstractBuffers for those results:
If the format of the results is known when creating the query, RWDBOSql may gain performance if output RWDBAbstractBuffers are bound to the result set before invoking execute(). However, binding output RWDBAbstractBuffers before invoking execute() is not required.
If the existence or format of the results is unknown--for example, if you are entering ad hoc queries--you would do the following:
Execute the RWDBOSql, and use methods such as hasResult() and schema() to discover the number and size of the result sets
Bind RWDBAbstractBuffers, such as RWDBTBuffer instances, to the result sets using the RWDBOSql operator[] and the RWDBMultiRow operator>>
Invoke fetch() to begin fetching results.
RWDBOSql is designed around the Interface/Implementation paradigm. An RWDBInserter instance is an interface to a reference-counted implementation; copy constructors and assignment operators produce additional references to a shared implementation.
In this example, we use an RWDBOSql to execute a query with a placeholder. The input value and output storage are both specified using RWDBTBuffer. Please note that Oracle-specific placeholder syntax is used here; check your access library guide or vendor documentation for appropriate SQL and placeholder syntax.
// Grab a connection from our database
RWDBConnection aConn = aDB.connection();
// Create the statement
RWDBOSql myQuery("SELECT TITLE,
TEMPO FROM SONGS WHERE KEY=:phKey");
// Bind in data
RWCString key;
RWDBTBuffer<RWCString> keyBinding(&key, 1);
myQuery << keyBinding;
// Set the query data; we're looking for songs in Bb
key = "Bb";
// Execute the query
myQuery.execute(aConn);
// Fetch and print results
if( myQuery.isValid() ) {
RWDBTBuffer<RWCString> titles(10);
RWDBTBuffer<unsigned> tempos(10);
// Bind result buffers to 0th result set
myQuery[0] >> titles >> tempos;
// Fetch data
while(myQuery.fetch(), myQuery.rowsFetched() > 0) {
// Print the rows that were fetched
for(int i=0; i<myQuery.rowsFetched(); i++ ) {
cout << "Song Title: " << titles[i]
<< " Tempo: " << tempos[i] << endl;
}
}
}
else
{
// An error
cout << "Error occurred: "
<< myQuery.status().message()
<< endl;
}
In this example, we use an RWDBOSql to execute a user-supplied ad hoc query. We then use an RWDBMultiRow to fetch the results.
// Utility function to print an RWDBRow
void outputRow(ostream &o, RWDBRow &r)
{
for(int i = 0; i < r.entries(); i++) {
if(i>0) cout << `\t';
RWDBValue theField = r[i];
cout << theField.asString();
}
cout << endl;
}
// This function is called by the user interface
void executeAndDisplayResults(RWDBConnection &c,
const RWCString &query)
{
// Create an RWDBOSql for the query
RWDBOSql osql(query);
// Execute the RWDBOSql
// No error checking performed here for simplicity
osql.execute(c);
// We'll use RWDBMultiRow to automatically create
// RWDBTBuffers for us from the schema of the results.
// The `1' indicates a one-row MultiRow - each
// RWDBTBuffer created will have one entry.
RWDBMultiRow outputBuffers( osql.schema(), 1 );
// Bind the MultiRow to the 0th result set
osql[0] = outputBuffers;
// Fetch and output the results
while( osql.fetch(), osql.rowsFetched() > 0 ) {
// Create an RWDBRow from the 0th row in the
// output buffers. Since we fetch only one row
// at a time,the row is always placed
// in the 0th row of the RWDBMultiRow.
RWDBRow aRow = outputBuffers[0];
// Print out the row we fetched
outputRow(cout, aRow);
}
}
enum StatementType { NonQuery = 0, Query, Procedure };
A class-scoped enum containing all the statement identifications understood by open SQL.
RWDBOSql();
Default constructor. Creates an RWDBOSql object with statement type RWDBOSql::NonQuery and no statement set. Its error handler is set to the error handler of the static RWDBManager.
RWDBOSql(const RWCString & sql,
StatementType type = NonQuery);
Creates an RWDBOSql with SQL statement set to sql, and statement type set to type. The statement type communicates to the access library the kind of query inside the statement. This information is used to help optimize binding and performance. Error handler for the created object is set to the error handler of the static RWDBManager, as in the default constructor.
RWDBOSql(const RWDBOSql& osql);
Creates an RWDBOSql whose implementation is shared with osql. All attributes are shared, including statement, status, and bindings.
~RWDBOSql();
Cleans up the RWDBOSql implementation if it is the last RWDBOSql to reference that implementation, consistent with standard Interface/Implementation destructor behavior. Blocks if an asynchronous operation has not completed. If blocking is not desired, call to cancel it before the RWDBOSql is destroyed.
RWDBOSql& operator=(const RWDBOSql& osql);
Assignment operator. Causes self's current implementation to be discarded, including its SQL statement, bindings, state, status, and any other attributes. Self will share osql's implementation.
RWDBOSql& operator<<(RWDBAbstractBuffer& input);
Sets an input binding on the current binding location. Advances the current binding location internally by one, so that the next call to operator<< binds the next binding location. Returns a reference to self.
RWDBMultiRow& operator[](size_t index);
Returns a reference to the RWDBMultiRow that stores all the output bindings for the indexth result set.
RWDBOSql& cancel(CancelType tp);
Attempts to cancel the pending result set, or current statement execution. The exact definition is vendor-dependent; see your access library guide for details.
NOTE:This call is only valid after the RWDBOSql is executed. Calling cancel() before execute() invalidates the RWDBOSql, setting an RWDBStatus::invalidUsage error, and calling any associated error handler.
RWDBStatus::errorHandler errorHandler() const;
Returns the error handler attached to self.
void execute(RWDBConnection& cn, size_t blkSz=0);
Executes the SQL statement on cn. If blkSz is nonzero, blkSz rows are sent to the server from the input bindings. If blkSz is zero, the number of rows sent is determined by the minimum of the lengths of all input-bound buffers.
RWDBOSql& fetch();
Fetches the current result set into the output bindings for that result set, and returns a reference to self. If the current result set is completely fetched, no rows are fetched, and the next call to fetch() moves to the next result set and fetches. The member function fetch() returns at most n rows at a time, where n is the minimum of the lengths of the output-bound RWDBTBuffers for the current result set. After calling fetch(), use rowsFetched() to determine how many rows were fetched.
NOTE:This call is only valid after the RWDBOSql is executed. Calling fetch() before execute() invalidates the RWDBOSql, setting an RWDBStatus::invalidUsage error, and calling any associated error handler.
RWDBOSql& fetchReturnParams();
Fetches any in/out or out-type stored procedure parameters into the bound buffers for those parameters. Returns a reference to self.
RWBoolean hasResult() const;
Returns TRUE if there is data waiting to be fetched. Also returns TRUE if all the data has been fetched, but the end-of-fetch is not yet encountered. If no attempt is made to read past the end of the result set that was fetched, hasResult() still returns TRUE.
RWBoolean isReady() const;
Returns TRUE if the object is in ready state, indicating that accessing the object will not block. Accessing a nonready object may potentially block.
RWBoolean isValid() const;
Returns TRUE if the RWDBOSql is not in an error condition. This call blocks if the last operation on this RWDBOSql is not yet completed.
RWDBValue returnValue() const;
Obtains and returns the stored procedure return value, if one exists. If no return value exists, or a stored procedure has not been executed, returns an RWDBValue whose type is RWDBValue::NoType.
long rowCount() const;
Returns the number of rows affected by the last execute(), if available. Returns -1 if this information is unavailable.
long rowsFetched() const;
Returns the number of rows fetched in the last fetch() operation. Blocks if the fetch() was asynchronous and has not yet completed. Returns -1 if no fetch was performed.
RWDBSchema schema() const;
Returns the schema for the current result set, if the statement is executed and has results. Returns an empty schema otherwise. This method is useful when the format of results are unknown, as is the case when using ad hoc queries.
void setErrorHandler(RWDBStatus::ErrorHandler err);
Installs err as self's error handler.
RWCString statement() const;
Returns the current SQL statement by value. If no statement is set, returns an empty string ("").
RWDBOSql&
statement(const RWCString& sql,
StatementType type = NonQuery);
Sets the statement to sql, and the statement type to type. All pending queries are canceled. All bindings are cleared. Any error set on the RWDBOSql is cleared.
RWDBStatus status() const;
Returns the status of self.
StatementType type() const;
Returns the statement type of the RWDBOSql. StatementType is a class-scoped enum containing all the statement identifications understood by open SQL.
©Copyright 1999, Rogue Wave Software, Inc.
Contact Rogue Wave about documentation or support issues.