Connect to an Oracle database. All subsequent operations will affect this database until the next call to ORACLE:CONNECT. A single program can access different Oracle schemas concurrently by repeated calls to ORACLE:CONNECT. Database connections are cached and re-used: if you call ORACLE:CONNECT again with the same user, schema, and server, the previous Oracle connection will be re-used. ORACLE:CONNECT may not be called inside WITH-TRANSACTION. Returns: T if a cached connection was re-used, NIL if a new connection was created (and cached). The meaning of the arguments is as follows:

Execute a SQL statement. Must be ORACLE:CONNECTed to a database. Returns the number of rows affected by the SQL operation, for non-SELECT statements, zero for SELECT statements. For destructive database operations (INSERT, UPDATE, DELETE), the results are committed to the database immediately if auto-commit when establishing the current connection; see ORACLE:CONNECT. The meaning of the arguments is as follows:

Macro which loops over a SQL SELECT result, evaluating, for each row in the result, the forms in body, binding symbols given in vars to corresponding database columns in the SELECT result. The argument vars must be a non-empty list of symbols matching a subset of the columns of an active SELECT query. If a SELECT column is an Oracle expression such as SUBSTR(mycol, 1, 10), it is recommended to use a column alias, e.g., SELECT SUBSTR(mycol, 1, 10) AS myvar, in which case the column alias will be used as the symbol bound to the column value.

As DO-ROWS expands into a DO* loop, it may be terminated prematurely, before all rows are fetched, by using RETURN anywhere in body.

It is allowed to call ORACLE:CONNECT in the body of the loop, but only to switch the connection to a database other than the one that was used to do the SELECT. This is useful for reading from one database while writing to another.

In vars, instead of a single symbol, a pair (bound-var "column-name") may be specified, which will cause values from the SELECTed column or alias , column-name, to be bound to Lisp variable, bound-var. This is for unusual cases where a Lisp variable cannot be created with the same name as the column (e.g., a column named "T"), or when it is inconvenient or impossible to alias the column with SELECT ... AS.

Fetch a single row of data. Returns a row of values corresponding to the columns of an active SELECT statment. The row data is returned in one of three different forms, depending on the value of the symbol result-type:

The following data type conversions are done between Oracle datatypes and Common Lisp data types:

Fetch some or all the rows from a query and return result as a sequence of sequences. Arguments are all optional: max-rows limits the result to that numbers of rows; result-type is the type of sequence of the rows, either 'ARRAY (the default) or 'LIST; item-type is the type of sequence of the column values for each row, either 'ARRAY (the default) or 'LIST. Each row fetched always contains the full set of column values SELECTed.

FETCH-ALL is often useful in conjunction with MAP or REDUCE to iterate over an entire SELECT result to construct a single Lisp value.

Returns information on the columns of a SELECT result, in the form of an array of SQLCOL structures, one for each result column in the most recent SELECT statement. It is not necessary to have called FETCH before requesting column information on the query, however the query must have been compiled and executed with RUN-SQL. Each SQLCOL structure has these slots:

To access the values of the SQLCOL structures, use the standard accessor functions, e.g., (ORACLE:SQLCOL-NAME (elt (ORACLE:COLUMNS) 0))