Description

The SQL Analyzer provides support for database technologies using the ANSI SQL-92 / 99 language. This extension uses the Universal Analyzer framework and is intended to analyse DDL, DML and SQL exports for a large variety of SQL variants:

• This extension provides source code analysis support for DDL and DML*.sql files using an over language of the various sql variants.
• This extension also accepts src and uaxDirectory files. Check here for more details about sqltablesize files.

In what situation should you install this extension?

• If you need to analyze PostgreSQL, MySQL, MariaDB, SQLite, DB2, Sybase, Microsoft SQL Server, Teradata or Informix
  
• If you need to analyze Oracle for the first time
• If your application contains schemas from database vendors not supported "out of the box" by CAST AIP but, which are compliant with ANSI SQL-92 / 99
• When you do not have access to the online database to perform an extraction for use with CAST AIP and have instead been provided with DDL scripts

Transitioning from from the CAST AIP DB2 Analyzer to the SQL Analyzer extension

If you have been actively analyzing DB2 (z/OS or UDB) with the DB2 Analyzer (provided out-of-the-box in CAST AIP) you can transition to using the SQL Analyzer extension to analyze your DB2 source code. The process of transitioning is described in SQL Analyzer - To do transition from the CAST AIP Db2 Analyzer to the SQL Analyzer extension.

Vendor compatibility matrix - official support

Icon	Vendor	Version	Supported?
	IBM DB2 - UDB/zOS	UDB: Up to version 11.1 z/OS: Up to version 12
	IBM Informix	Up to version 12.x
	MariaDB	Up to version 10.x
	Microsoft SQL Server	Up to version 2019
	MySQL	Up to version 5.5
	Oracle Server	Up to version 12c	(See note below)
	PostgreSQL	Up to version 11
	SAP ASE (formerly known as Sybase ASE)	Up to version 16
	SQLite	Up to version 3.x
	Teradata	Up to version 16

Function Point, Quality and Sizing support

This extension provides the following support:

• Function Points (transactions): a green tick indicates that OMG Function Point counting and Transaction Risk Index are supported
• Quality and Sizing: a green tick indicates that CAST can measure size and that a minimum set of Quality Rules exist

Function Points (transactions)	Quality and Sizing

CAST AIP compatibility

This extension is compatible with:

CAST AIP release	Supported
8.3.x
8.2.x
8.1.x
8.0.x
7.3.x

Supported DBMS servers used for CAST AIP schemas

This extension is compatible with the following DBMS servers used to host CAST AIP schemas:

CAST AIP release	CSS	Oracle	Microsoft
All supported releases

Prerequisites

An installation of any compatible release of CAST AIP (see table above)

Download and installation instructions

Please see:

• Download an extension
• Install an extension

Upgrade from the SQL Script extension

If you have previously used the SQL Script extension (com.castsoftware.sqlscript) on existing schemas, you should proceed as following:

• In CAST Server Manager use the Manage Extensions option on the CAST AIP schemas in which the SQL Script extension is installed
• Select Analyzer for SQL files and choose deactivate to remove the existing extension. No further actions are required.

Packaging, delivering and analyzing your source code

Please see: SQL Analyzer - 2.x and 3.x - Packaging, delivering and analyzing your source code.

What results can you expect?

Once the analysis/snapshot generation has completed, you can view the results in the normal manner (for example via CAST Enlighten) - click to enlarge:

You can also use the CAST Management Studio option View Analysis Unit Content to see the objects that have been created following the analysis:

Objects

The following objects are displayed in CAST Enlighten:

Table deletion and renaming

DROP TABLE syntax is supported for table objects within the same file. When creating a table through CREATE TABLE tableName (colName int, ...) followed by a DROP TABLE tableName, the table will not be recorded and thus will not be displayed in CAST Enlighten. Similarly, if a table is renamed with a RENAME TABLE statement (or ALTER TABLE RENAME TO as in SQLite and PostgreSQL), this change will be reflected in CAST Enlighten. Presently we consider case-insensitive names, i.e., objects named tableName, TABLEname are considered to be the same object.

Database objects for DB2 analyses

Please read this note if you move from sqlanalyzer <= 2.6.9-funcrel to sqlanalyzer >= 3.0.0-alpha2 and if the AIP core is from 8.3.16 to 8.3.28 and analyzed sources are DB2:

The following CREATE TABLE ... IN ... statements will generate a new object: a database for DB2 analyses:

CREATE TABLE .... () ... IN DATABASE_NAME.TABLESPACE_NAME ....  --see ABC 02 example

CREATE TABLE .... () IN DATABASE DATABASE_NAME --see ABC 03 example

Object Full Name will change from <SCHEMA NAME>.<OBJECT NAME> to <DATABASE NAME>.<SCHEMA NAME>.<OBJECT NAME:

Links

Links are created for transaction and function point needs:

DDL

You can expect the following links on the DDL side within the same sql file:

• useSelect, useInsert, useUpdate, useDelete Links from Procedure / Function / Event  to Table / View
• callLink from Procedure / Function / Event to Procedure / Function
• callLink from Procedure / Function  to Cobol Program

• callLink from Procedure / Function  to C/C++ Function

• callLink from Procedure / Function  to Java Method

• useSelect from View to Table / View used in the query of the view
• callLink from View to Function
• relyonLink from Index to the Table
• relyonLink from Index to the Column implied in the index
• relyonLink from Synonym to Table / View / Function / Procedure / Package aliased by Synonym
• referLink from Table / Table Column to a Table / Table Column referenced in a Foreign Key 
• referDelete, referUpdate from Table to a Table referenced in a Foreign Key. Examples:
  
  • referDelete:

ALTER TABLE A.TABLE_1 ADD CONSTRAINT FK_NAME_1 FOREIGN KEY (COL_1) REFERENCES A.TABLE_2 (COL_2) ON DELETE;

• • referUpdate:

ALTER TABLE A.TABLE_1 ADD CONSTRAINT FK_NAME_1 FOREIGN KEY (COL_1) REFERENCES A.TABLE_2 (COL_2) ON UPDATE;

• • referDelete with a Referential Action property Restrict for Delete: 

ALTER TABLE A.TABLE_1 ADD CONSTRAINT FK_NAME_1 FOREIGN KEY (COL_1) REFERENCES A.TABLE_2 (COL_2) ON DELETE RESTRICT;

• • referDelete with a Referential Action property Cascade for Delete : 

ALTER TABLE A.TABLE_1 ADD CONSTRAINT FK_NAME_1 FOREIGN KEY (COL_1) REFERENCES A.TABLE_2 (COL_2) ON DELETE CASCADE;

• • referDelete with a Referential Action property Cascade for Delete , referUpdate with a Referential Action property No Action for Update

CREATE TABLE IF NOT EXISTS `TABLE_1` (
  `COL_1` int NOT NULL,
  `COL_2` char(4) COLLATE latin1_bin NOT NULL,
  KEY `FK__TABLE_1__TABLE_2___5AEE82B9` (`COL_1`),
  CONSTRAINT `FK__TABLE_1__TABLE_2___5AEE82B9` FOREIGN KEY (`COL_1`) REFERENCES `TABLE_1` (`COL_1`) ON DELETE CASCADE ON UPDATE NO ACTION
) ENGINE=InnoDB DEFAULT CHARSET=latin1 COLLATE=latin1_bin;

• • referDelete with a Referential Action property No Action for Delete , referUpdate with a Referential Action property No Action for Update

• • referDelete with a Referential Action property Cascade for Delete 

• callLink to the correct Trigger where the tables is accessed in insert/update/delete
  • example a Trigger declared as BEFORE INSERT on a table, any insert to that table will call the trigger...

• example: a table, with an after update trigger, updated in a Client code :

• inheritLink from sub Type to super Type:
  
  • example : CREATE OR REPLACE type gayerrortype under gayxmlmsg ....

• useLink for PL/SQL tables, example : CREATE OR REPLACE TYPE "CNCUBLIST_2" is table of "HIS_TEST_A" ;

DML

You can expect the following links on the DML side :

• useSelect, useInsert, useUpdate, useDelete Links from SQL Script to Table / View
• call Links from SQL Script  to Procedure / Function

Rules

You can find a full list of rules delivered with this extension here:

Client Side Support: COBOL, PB, VB, .NET, JAVA, C/C++, PYTHON

Special note about XXL/XXS support

See SQL Analyzer - working with XXL or XXS tables for more information.

Special notes about Quality Rules on client side

Some Quality Rules are calculated on SQL queries on the client-side with some limitations:

• Quality Rules will be enabled on client-side code only if the server-side code has been analyzed with SQL Analyzer extension.
• For Java client-side code, SQL statements used in parameters of methods including a SQL parametrization rule are analyzed.

class Foo
{
   final static String TABLE_NAME = "Person";

	void method()
	{
    	String query = "select * from " + this.TABLE_NAME;
    	java.sql.Statement.execute(query );
	}
}

• But 'queries' visible in the DLM (that need reviewing) are not analyzed:

class Foo
{
	// not passed to an execute something
	private final static String text = "select name from Person";
}

• Explicit queries used in an ORM context are analyzed (or not) based on if they are visible in Enlighten

• COBOL EXEC SQL queries are analyzed

• SQL queries founded in Python code

• SQL queries founded in .properties (Java Property Mapping objects) are analyzed

Special note about redundant Quality Rules

Please see SQL Analyzer - Redundant Quality Rules not included in the SQL Analyzer.

Errors and warning

Please see SQL Analyzer - errors and warnings for the full list of analysis errors and warnings.

Known limitations and issues

Installation

If you encounter the following error in CAST Server Manager while installing the SQL Analyzer extension, please perform the workaround described here and then attempt the installation again. This error may occur if you have installed a very old and unsupported custom Universal Analyzer language pack that used the same metamodel type names as used in the official SQL Analyzer extension.

SQL Analyzer is incompatible with the schema metamodel. It is generally due to an extension that has changed its ids.

Analysis

• All name resolution is considered as case insensitive: this may produce erroneous links on a case insensitive platform 'playing with case': two different tables with the same case insensitive name will be both called
• Procedure resolution handles overriding when the number of parameters are matched or number and optionals are matched. Otherwise, when calling an overridden procedure, all overrides will be called. Below are some examples here is a single call Link, between the second func1 and func2:

CREATE FUNCTION func1() RETURNS integer AS
begin
    DELETE FROM table1 WHERE ID in (SELECT ID FROM table2);
end;

CREATE FUNCTION func1(mode integer) RETURNS integer AS
begin
    DELETE FROM table1 WHERE ID (SELECT ID FROM table2);
end;

CREATE FUNCTION func2(mode integer) RETURNS integer AS
begin
    select func1 (mode); 
end;

CREATE FUNCTION func1(i_mode integer) RETURNS integer AS
begin
    DELETE FROM table1 WHERE ID in (SELECT ID FROM table2);
end;

CREATE FUNCTION func1(mode integer := 1) RETURNS integer AS
begin
    DELETE FROM table1 WHERE ID in (SELECT ID FROM table2);
end;

CREATE FUNCTION func2(mode integer) RETURNS integer AS
begin
    select func1 (); 
end;

• Dynamic SQL statements are resolved when:
  • the SQL statement is readable, even for sliced statements.

  • TABLE1, TABLE2, TABLE3 and TABLE4 are visibles and useSelect link were be added

CREATE PROCEDURE test
AS
L_QryStr varchar2(4000);
begin
L_QryStr := 'select S.COL1, P.COL2, P.COL3, P.COL4, P.COL5, ' ||
' P.COL6, B.COL7, R.COL8, B.COL9, B.COL10' ||
' from TABLE1 S, TABLE2 P, TABLE3 B, ' ||
' ( select distinct R.COL1, R.COL2 ' ||
' from TABLE4 R ' ||
' where R.COL3 = 99999 ';
EXECUTE IMMEDIATE L_QryStr;
end;
/

• • test_table is visible and useDelete link is added:

CREATE PROCEDURE test
AS
begin
EXECUTE IMMEDIATE 'truncate table test_table'; 
end;
/

• • test_table is visible and useDelete link is added:

CREATE PROCEDURE test
AS
L_QryStr varchar2(4000);
begin
	L_QryStr := 'truncate table ';
	L_QryStr := L_QryStr || ' test_table ';
	EXECUTE IMMEDIATE L_QryStr; 
end;
/

• • table name is valued via a variable which could be resolved.
  • emp table is visible and useSelect link will be added

CREATE PROCEDURE test
AS
emp_rec  emp%ROWTYPE;
sql_stmt VARCHAR2(200);
begin
    sql_stmt := 'SELECT * FROM || emp_rec.T1 ||  WHERE job = 1';
    EXECUTE IMMEDIATE sql_stmt;
end;
/

• • emp table is visible and useSelect link will be added

CREATE OR REPLACE package body test_package as
 
 type T_1 is table of varchar2(22);
 emp_rec  emp%ROWTYPE;

 PROCEDURE test
is
sql_stmt VARCHAR2(2000);
emp_tab VARCHAR2(200);
begin
    emp_tab := emp_rec;
    sql_stmt := 'SELECT * FROM || emp_tab ||  WHERE job = :j';
    EXECUTE IMMEDIATE sql_stmt;
end test;
 
end;
/

• • OPEN-FOR-USING : emp table and test procedure are linked by a use SELECT link

create table emp (col1 int);
/
CREATE OR REPLACE type body test_body as

 type t_emp is table of t_table index by varchar2(22);
 emp_rec  emp%ROWTYPE;

 member PROCEDURE test(o_cursor OUT my_cursor)
IS
BEGIN
   OPEN    o_cursor FOR
   'SELECT  rule_id,
           expression_id,
           parent_expression_id,
           operator
   FROM    emp
   ORDER   BY rule_id, expression_id';
END test;

end;
/

create table emp (col1 int);
/
CREATE OR REPLACE type body test_body as

 type t_emp is table of t_table index by varchar2(22);
 emp_rec  emp%ROWTYPE;

 member PROCEDURE test(o_cursor OUT my_cursor)
IS
L_SQL varchar(20000) := '';
BEGIN
    L_SQL := 'SELECT  rule_id,
           expression_id,
           parent_expression_id,
           operator
   FROM    emp
   ORDER   BY rule_id, expression_id';
   
   OPEN o_cursor FOR L_SQL;
END test;

end;
/

• ALTER TABLE ... ADD ... syntax is supported. All other syntaxes, such as ALTER TABLE ... DELETE .. or ALTER TABLE ... DROP ... or ALTER TABLE ... MODIFY ... etc. are not supported.

• Moving a table from one database/scheme to another is not supported through RENAME TABLE schema1.tableName1 TO schema2.tableName2.   
• Sequences are not taken in to account and that is not a limitation but a choice because they have no effect on transactions nor Quality Rules
• Oracle synonyms on packages are not taken in to account.
• For the QR 1101012 Specify column names instead of column numbers in ORDER BY clauses, the case when a function that returns a number or a numeric variable is used in order by is not reported to violate the rule.

CAST Health Dashboard (ex. Application Analytics Dashboard)

Starting with release 2.1 of the SQL Analyzer extension, the name used to represent the technology has changed from SQL Analyzer to SQL: if you have already transfered snapshots that contain SQL Analyzer analysis results in to a Measurement Service schema, you'll see SQL Analyzer in the list of technologies instead of SQL.

If you would like to change the technology name for existing snapshots, you can change it using the following SQL query run against the Measurement Service schema:

update DSS_OBJECT_TYPES
set OBJECT_TYPE_NAME = 'SQL'
where OBJECT_TYPE_ID = 1101000