Archive for January, 2015
A PL/pgSQL Function
Somebody wanted to know how to write a basic PostgreSQL PL/pgSQL function that returned a full name whether or not the middle name was provided. That’s pretty simple. There are principally two ways to write that type of concatenation function. One uses formal parameter names and the other uses positional values in lieu of the formal parameter names.
The two ways enjoy two techniques (SQL language and PL/pgSQL language), which gives us four possible solutions. I’ve also provided a conditional drop statement for the full_name function. If you’re new to PostgreSQL the DROP
statement might make you scratch your head because you’re wondering why you need to use the formal parameter list. The DROP
statement needs the parameter list because PostgeSQL lets you overload schema/database functions and procedures.
The code is for a named parameter lists using the SQL language is:
DROP FUNCTION IF EXISTS full_name ( IN pv_first_name text , IN pv_middle_name text , IN pv_full_name text); CREATE FUNCTION full_name ( IN pv_first_name text , IN pv_middle_name text , IN pv_last_name text , OUT pv_full_name text) AS 'SELECT pv_first_name || CASE WHEN pv_middle_name IS NOT NULL THEN '' '' || pv_middle_name || '' '' ELSE '' '' END || pv_last_name' LANGUAGE SQL; |
The code is for a positional parameter lists using the SQL language is:
DROP FUNCTION IF EXISTS full_name ( IN text , IN text , IN text); CREATE FUNCTION full_name ( IN text , IN text , IN text , OUT text) AS 'SELECT $1 || CASE WHEN $2 IS NOT NULL THEN '' '' || $2 || '' '' ELSE '' '' END || $3' LANGUAGE SQL; |
You would re-write the function in the PL/pgSQL language as follows (please note the named parameter list):
CREATE FUNCTION full_name ( IN pv_first_name text , IN pv_middle_name text , IN pv_last_name text) RETURNS text AS $$ DECLARE lv_output text; BEGIN IF pv_middle_name IS NULL THEN lv_output = CONCAT(pv_first_name, N' ', pv_last_name); ELSE lv_output = CONCAT(pv_first_name, N' ', pv_middle_name, N' ', pv_first_name); END IF; RETURN lv_output; END $$ LANGUAGE plpgsql IMMUTABLE; |
You can test either version of the program with the following two queries from the pseudo table dual, which isn’t require in the SELECT
statement:
SELECT full_name('Henry',NULL,'Pym') AS "Ant-Man" UNION ALL SELECT full_name('Henry','''Hank''','Pym') AS "Ant-Man"; |
It prints:
Ant-Man text ---------------- Henry Pym Henry 'Hank' Pym |
As always, I hope this helps those looking for how to accomplish a concatenation function in PostgreSQL.
Convert to SQL Server?
I’m always amazed at the questions that pop up for me. For example, how do you convert an Oracle script that creates my Video Store model to a Microsoft SQL Server script. It’s not very hard but there’s one big caveat, and that’s the fact that system_user
is a reserved word. That means you can’t create the Access Control List (ACL) table with a system_user
name. The alternative, would be to convert the system_user
table name to database_user
. That’s what I’ve done in this example.
It’s also important to note that this example uses Microsoft SQL Server’s sqlcmd
in batch mode. Naturally, it presumes that you’ve created a student
user with a trivial password of student
, and a studentdb
schema. Also, that you’ve granted privileges so everything works (if you need help on that check my earlier post on how to setup a studentdb
schema).
The following is an example of conditionally dropping and then creating a system_user
table in an Oracle schema. It uses a CASCADE CONSTRAINTS
clause to eliminate dependencies with foreign key values.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 | -- Conditionally drop the table and sequence. BEGIN FOR i IN (SELECT NULL FROM user_tables WHERE table_name = 'SYSTEM_USER') LOOP EXECUTE IMMEDIATE 'DROP TABLE system_user CASCADE CONSTRAINTS'; END LOOP; FOR i IN (SELECT NULL FROM user_sequences WHERE sequence_name = 'SYSTEM_USER_S1') LOOP EXECUTE IMMEDIATE 'DROP SEQUENCE system_user_s1'; END LOOP; END; / |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 | -- Create SYSTEM_USER table. CREATE TABLE system_user ( system_user_id NUMBER CONSTRAINT system_user_pk PRIMARY KEY , system_user_name VARCHAR2(20) CONSTRAINT system_user_nn1 NOT NULL , system_user_group_id NUMBER CONSTRAINT system_user_nn2 NOT NULL , system_user_type NUMBER CONSTRAINT system_user_nn3 NOT NULL , first_name VARCHAR2(20) , middle_name VARCHAR2(20) , last_name VARCHAR2(20) , created_by NUMBER CONSTRAINT system_user_nn4 NOT NULL , creation_date DATE CONSTRAINT system_user_nn5 NOT NULL , last_updated_by NUMBER CONSTRAINT system_user_nn6 NOT NULL , last_update_date DATE CONSTRAINT system_user_nn7 NOT NULL , CONSTRAINT system_user_fk1 FOREIGN KEY (created_by) REFERENCES system_user (system_user_id) , CONSTRAINT system_user_fk2 FOREIGN KEY (last_updated_by) REFERENCES system_user (system_user_id)); -- Create SYSTEM_USER_S1 sequence with a start value of 1001. CREATE SEQUENCE system_user_s1 START WITH 1001; |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 | -- Conditionally drop the table and sequence. BEGIN FOR i IN (SELECT NULL FROM user_tables WHERE table_name = 'COMMON_LOOKUP') LOOP EXECUTE IMMEDIATE 'DROP TABLE common_lookup CASCADE CONSTRAINTS'; END LOOP; FOR i IN (SELECT NULL FROM user_sequences WHERE sequence_name = 'COMMON_LOOKUP_S1') LOOP EXECUTE IMMEDIATE 'DROP SEQUENCE common_lookup_s1'; END LOOP; END; / |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | -- Create COMMON_LOOKUP table. CREATE TABLE common_lookup ( common_lookup_id NUMBER , common_lookup_context VARCHAR2(30) CONSTRAINT nn_clookup_1 NOT NULL , common_lookup_type VARCHAR2(30) CONSTRAINT nn_clookup_2 NOT NULL , common_lookup_meaning VARCHAR2(30) CONSTRAINT nn_clookup_3 NOT NULL , created_by NUMBER CONSTRAINT nn_clookup_4 NOT NULL , creation_date DATE CONSTRAINT nn_clookup_5 NOT NULL , last_updated_by NUMBER CONSTRAINT nn_clookup_6 NOT NULL , last_update_date DATE CONSTRAINT nn_clookup_7 NOT NULL , CONSTRAINT pk_c_lookup_1 PRIMARY KEY(common_lookup_id) , CONSTRAINT fk_c_lookup_1 FOREIGN KEY(created_by) REFERENCES system_user(system_user_id) , CONSTRAINT fk_c_lookup_2 FOREIGN KEY(last_updated_by) REFERENCES system_user(system_user_id)); -- Create a non-unique index on a single column. CREATE INDEX common_lookup_n1 ON common_lookup(common_lookup_context); -- Create a unique index based on two columns. CREATE UNIQUE INDEX common_lookup_u2 ON common_lookup(common_lookup_context,common_lookup_type); -- Create COMMON_LOOKUP_S1 sequence with a start value of 1001. CREATE SEQUENCE common_lookup_s1 START WITH 1001; |
You can do the same thing for a database_user
table in Microsoft SQL Server with the following syntax. Unfortunately, there isn’t a CASCADE CONSTRAINTS
clause that we can append in Microsoft SQL Server. The script uses a dynamic SQL statement with a Common Table Expression (CTE) to generate a list of ALTER
statements that drop foreign key constraints in the schema.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 | /* Drop all foreign keys. */ USE studentdb; /* Create a session variable to hold a command list. */ SELECT 'Create a session variable.' AS "Statement"; DECLARE @sql NVARCHAR(MAX) = N''; /* Generate the command list to drop foreign key constraints. */ SELECT 'Generate dynamic SQL statements.' AS "Statement"; ;WITH x AS (SELECT N'ALTER TABLE ' + OBJECT_SCHEMA_NAME(parent_object_id) + N'.' + OBJECT_NAME(parent_object_id) + N' ' + N'DROP CONSTRAINT ' + name + N';' AS sqlstmt FROM sys.foreign_keys) SELECT @sql += sqlstmt FROM x; /* Call the dynamically generated statements. */ SELECT 'Execute dynamic SQL statements.' AS "Statement"; EXEC sp_executesql @sql; /* Conditionally drop tables. */ SELECT 'Conditionally drop studentdb.common_lookup table.' AS "Statement"; IF OBJECT_ID('studentdb.database_user','U') IS NOT NULL DROP TABLE studentdb.database_user; /* Create a table with self-referencing foreign key constraints. */ SELECT 'Create studentdb.common_lookup table.' AS "Statement"; CREATE TABLE studentdb.database_user ( database_user_id INT NOT NULL IDENTITY(1,1) CONSTRAINT database_user_pk PRIMARY KEY , database_user_name VARCHAR(20) NOT NULL , database_user_group_id INT NOT NULL , database_user_type INT NOT NULL , first_name VARCHAR(20) , middle_name VARCHAR(20) , last_name VARCHAR(20) , created_by INT NOT NULL , creation_date DATE NOT NULL , last_updated_by INT NOT NULL , last_update_date DATE NOT NULL , CONSTRAINT database_user_fk1 FOREIGN KEY (created_by) REFERENCES studentdb.database_user (database_user_id) , CONSTRAINT database_user_fk2 FOREIGN KEY (created_by) REFERENCES studentdb.database_user (database_user_id)); /* Conditionally drop common_lookup table. */ SELECT 'Conditionally drop studentdb.common_lookup table.' AS "Statement"; IF OBJECT_ID('studentdb.common_lookup','U') IS NOT NULL DROP TABLE studentdb.common_lookup; /* Create a table with external referencing foreign key constraints. */ SELECT 'Create studentdb.common_lookup table.' AS "Statement"; CREATE TABLE studentdb.common_lookup ( common_lookup_id INT NOT NULL IDENTITY(1,1) CONSTRAINT common_lookup_pk PRIMARY KEY , common_lookup_context VARCHAR(30) CONSTRAINT nn_clookup_1 NOT NULL , common_lookup_type VARCHAR(30) CONSTRAINT nn_clookup_2 NOT NULL , common_lookup_meaning VARCHAR(30) CONSTRAINT nn_clookup_3 NOT NULL , created_by INT CONSTRAINT nn_clookup_4 NOT NULL , creation_date DATE CONSTRAINT nn_clookup_5 NOT NULL , last_updated_by INT CONSTRAINT nn_clookup_6 NOT NULL , last_update_date DATE CONSTRAINT nn_clookup_7 NOT NULL , CONSTRAINT common_lookup_fk1 FOREIGN KEY(created_by) REFERENCES studentdb.database_user (database_user_id) , CONSTRAINT common_lookup_fk2 FOREIGN KEY(last_updated_by) REFERENCES studentdb.database_user (database_user_id)); |
You can run it from a file by calling the sqlcmd
utility. You’ll need to know several things to run it. First, you need to know your database instance. You can capture that from a query against the data dictionary or catalog. Just run the following from inside the Microsoft SQL Server Management Studio (SSMS):
SELECT @@SERVERNAME; |
In my case, it shows the following, which is the machine’s hostname
a backslash and SQLEXPRESS
:
MCLAUGHLINSQL\SQLEXPRESS |
The script uses sqltest.sql
as a file name, and you can call it from the Windows shell environment like this:
sqlcmd -S MCLAUGHLINSQL\SQLEXPRESS -U student -P student -i C:\Data\MicrosoftSQL\sqltest.sql -o C:\Data\Microsoft\sqltest.out |
As always, I hope this helps.
Popular Programming Languages
First of all, Happy New Year!
IEEE Spectrum published a ranking of the most popular programming languages. Computational journalist Nick Diakopoulos wrote the article. While it may surprise some, I wasn’t surprised to find SQL in the top ten.
Nick weighted and combined 12 metrics from 10 sources (including IEEE Xplore, Google, and GitHub) to rank the most popular programming languages.
- Compiled programming languages (Java [#1], C [#2], C++ [#3], C# [#4], Objective-C [#16])
- Interpreted programming languages (Python [#5], JavaScript [#6], PHP [#7], Ruby [#8], Perl [#11], HTML [#12])
- Data languages (SQL [#9], MATLAB [#10], R [#13])
I couldn’t resist including Objective-C because it shows how the iPhone, iPad, and Mac OS impact our daily lives. At the same time, Assembly [#15] is actually more popular than Objective-C. Shell [#17] follows Objective-C. While the Visual Basic [#14] programming language still remains very popular.
There are many “why” questions raised by this list of popular programming languages. The “why” from my perspective deals with what are the market drivers for their popularity. The money drivers I see are as follows:
- Business software: Java, C++, C#, and AIDE – Android IDE (works with Java and C++ source code)
- OS X and iOS Development: Objective-C
- Development Tools: Java, C, C++, and Python
- System Admin/Utilities Tools: C, Perl, and Shell
- Web Development: Python, PHP, Ruby, and Perl
- Data Analysis: SQL, MATLAB, and R
Business Intelligence (BI) software manages most high-level data analysis tools and they’ll continue to get better over time. However, if SQL has shown us anything over 30 years it’s that ultimately we revert to it to solve problems. The conclusion from the reality of BI probably means the programming languages that develop those tools will continue to rise and so will the underlying data languages.
It’s also interesting to note that nine out of ten of the popular programming languages work with databases, like Oracle, MySQL, PostgreSQL, or SQL Server. While JavaScript doesn’t access the database typically, it’s JSON (JavaScript Object Notation) is supported in all the databases.