PostgreSQL Table Function
A quick tutorial on how to write a PL/pgSQL Table function. The functions is simple. It returns the list of conquistadors that were originally German. It does that by filtering on the lang column in the table. For example, you use ‘de‘ for German.
I’ll stage this with the same conquistador table used in the last post. Don’t forget to use the chcp command to the Active Console Code Page to 4-byte Unicode before you run the script file, like:
chcp 65001 |
Then, connect to the psql shell and run the following script file:
/* Conditionally drop the conquistador table. */ DROP TABLE IF EXISTS conquistador; /* Create the conquistador table. */ CREATE TABLE conquistador ( conquistador_id SERIAL , conquistador VARCHAR(30) , actual_name VARCHAR(30) , nationality VARCHAR(30) , lang VARCHAR(2)); /* Insert some conquistadors into the table. */ INSERT INTO conquistador ( conquistador , actual_name , nationality , lang ) VALUES ('Juan de Fuca','Ioánnis Fokás','Greek','el') ,('Nicolás de Federmán','Nikolaus Federmann','German','de') ,('Sebastián Caboto','Sebastiano Caboto','Venetian','it') ,('Jorge de la Espira','Georg von Speyer','German','de') ,('Eusebio Francisco Kino','Eusebius Franz Kühn','Italian','it') ,('Wenceslao Linck','Wenceslaus Linck','Bohemian','cs') ,('Fernando Consag','Ferdinand Konšcak','Croatian','sr') ,('Américo Vespucio','Amerigo Vespucci','Italian','it') ,('Alejo García','Aleixo Garcia','Portuguese','pt'); |
Now, you can build another script file to create the getConquistador function, like:
/* Drop the funciton conditionally. */ DROP FUNCTION IF EXISTS getConquistador; |
Create the getConquistador function:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | CREATE FUNCTION getConquistador (IN lang_in VARCHAR(2)) RETURNS TABLE ( conquistador VARCHAR(30) , actual_name VARCHAR(30) , nationality VARCHAR(30)) AS $$ BEGIN RETURN QUERY SELECT c.conquistador , c.actual_name , c.nationality FROM conquistador c WHERE c.lang = lang_in; END; $$ LANGUAGE plpgsql; |
Then, you can test it like:
SELECT * FROM getConquistador('de'); |
It will return the following:
conquistador | actual_name | nationality -----------------------+--------------------+------------- Nicolás de Federmán | Nikolaus Federmann | German Jorge de la Espira | Georg von Speyer | German (2 rows) |
As always, I hope this helps with a technique that’s useful.
PostgreSQL Unicode
It seems unavoidable to use Windows. Each time I’m compelled to run tests on the platform I find new errors. For example, they don’t use 4-byte unicode and as a result when you want to use Unicode in PostgreSQL there’s a mismatch.
For example, change the Active Console Code Page with the chcp (change code page) to match the one PostgreSQL uses, like:
chip 1252 |
It lets you avoid this warning message:
Password for user postgres:
psql (14.1)
WARNING: Console code page (437) differs from Windows code page (1252)
8-bit characters might not work correctly. See psql reference
page "Notes for Windows users" for details.
Type "help" for help.
postgres=# |
However, it won’t avoid display issues with real Unicode values. For example, let’s use a small international table like the following:
/* Conditionally drop the conquistador table. */ DROP TABLE IF EXISTS conquistador; /* Create the conquistador table. */ CREATE TABLE conquistador ( conquistador_id SERIAL , conquistador VARCHAR(30) , actual_name VARCHAR(30) , nationality VARCHAR(30) , lang VARCHAR(2)); /* Insert some conquistadors into the table. */ INSERT INTO conquistador ( conquistador , actual_name , nationality , lang ) VALUES ('Juan de Fuca','Ioánnis Fokás','Greek','el') ,('Nicolás de Federmán','Nikolaus Federmann','German','de') ,('Sebastián Caboto','Sebastiano Caboto','Venetian','it') ,('Jorge de la Espira','Georg von Speyer','German','de') ,('Eusebio Francisco Kino','Eusebius Franz Kühn','Italian','it') ,('Wenceslao Linck','Wenceslaus Linck','Bohemian','cs') ,('Fernando Consag','Ferdinand Konšcak','Croatian','sr') ,('Américo Vespucio','Amerigo Vespucci','Italian','it') ,('Alejo García','Aleixo Garcia','Portuguese','pt'); /* Query the values from the conquistador table. */ SELECT * FROM conquistador; |
When you call the script to load it, like:
\i testScript.sql |
It’ll display the following, which you can check against the strings in the VALUES clause above. There are encoding issues on lines 1, 2, 3, 5, 7, and 8 below.
conquistador_id | conquistador | actual_name | nationality | lang
-----------------+------------------------+----------------------+-------------+------
1 | Juan de Fuca | Ioánnis Fokás | Greek | el
2 | Nicolás de Federmán | Nikolaus Federmann | German | de
3 | Sebastián Caboto | Sebastiano Caboto | Venetian | it
4 | Jorge de la Espira | Georg von Speyer | German | de
5 | Eusebio Francisco Kino | Eusebius Franz Kühn | Italian | it
6 | Wenceslao Linck | Wenceslaus Linck | Bohemian | cs
7 | Fernando Consag | Ferdinand Konšcak | Croatian | sr
8 | Américo Vespucio | Amerigo Vespucci | Italian | it
9 | Alejo GarcÃa | Aleixo Garcia | Portuguese | pt
(9 rows) |
If you’re like me, it was annoying. The problem is that the native 2-byte Unicode of Microsoft sends values into PostgreSQL that are invalid. Those codes are read back with unintended values from other character encoding sets.
While you can’t set Windows generic encoding to 65001 without causing the system problems, you can set Active Console Code Page value in the scope of a Command-Line session before running the script.
The chcp command lets you set it to 4-byte Unicode, like:
chcp 65001 |
Now, rerun the script and PostgreSQL will display the correct character encoding set with some spacing irregularities. However, that’s not what’s important when you call table from another programming language through the ODBC-layer. The data will be returned in a 4-byte Unicode encoding stream.
conquistador_id | conquistador | actual_name | nationality | lang
-----------------+------------------------+----------------------+-------------+------
1 | Juan de Fuca | Ioánnis Fokás | Greek | el
2 | Nicolás de Federmán | Nikolaus Federmann | German | de
3 | Sebastián Caboto | Sebastiano Caboto | Venetian | it
4 | Jorge de la Espira | Georg von Speyer | German | de
5 | Eusebio Francisco Kino | Eusebius Franz Kühn | Italian | it
6 | Wenceslao Linck | Wenceslaus Linck | Bohemian | cs
7 | Fernando Consag | Ferdinand Konšcak | Croatian | sr
8 | Américo Vespucio | Amerigo Vespucci | Italian | it
9 | Alejo García | Aleixo Garcia | Portuguese | pt
(9 rows) |
A similar error to what I encountered testing MySQL Workbench’s ability to export SQL Server databases 10 years ago. I thought giving a solution to get coerce correct 4-byte Unicode data insertion may help those who also may be surprised by the behavior.
PL/pgSQL List to Struct
This blog post addresses how to convert a list of values into a structure (in C/C++ its a struct, in Java its an ArrayList, and PL/pgSQL it’s an array of a type). The cast_strings function converts a list of strings into a record data structure. It calls the verify_date function to identify a DATE data type and uses regular expressions to identify numbers and strings.
You need to build the struct type below first.
CREATE TYPE struct AS ( xnumber DECIMAL , xdate DATE , xstring VARCHAR(100)); |
The cast_strings function is defined below:
CREATE FUNCTION cast_strings ( pv_list VARCHAR(10)[] ) RETURNS struct AS $$ DECLARE /* Declare a UDT and initialize an empty struct variable. */ lv_retval STRUCT := (null, null, null); BEGIN /* Loop through list of values to find only the numbers. */ FOR i IN 1..ARRAY_LENGTH(pv_list,1) LOOP /* Order if statements by evaluation. */ CASE /* Check for a value with only digits. */ WHEN lv_retval.xnumber IS NULL AND REGEXP_MATCH(pv_list[i],'^[0-9]+$') IS NOT NULL THEN lv_retval.xnumber := pv_list[i]; /* Check for a valid date. */ WHEN lv_retval.xdate IS NULL AND verify_date(pv_list[i]) IS NOT NULL THEN lv_retval.xdate := pv_list[i]; /* Check for a string with characters, whitespace, and digits. */ WHEN lv_retval.xstring IS NULL AND REGEXP_MATCH(pv_list[i],'^[A-Za-z 0-9]+$') IS NOT NULL THEN lv_retval.xstring := pv_list[i]; ELSE NULL; END CASE; END LOOP; /* Print the results. */ RETURN lv_retval; END; $$ LANGUAGE plpgsql; |
There are two test cases for the cast_strings function. One uses a DO-block and the other a query.
- The first use-case checks with a DO-block:
DO $$ DECLARE lv_list VARCHAR(11)[] := ARRAY['86','1944-04-25','Happy']; lv_struct STRUCT; BEGIN /* Pass the array of strings and return a record type. */ lv_struct := cast_strings(lv_list); /* Print the elements returned. */ RAISE NOTICE '[%]', lv_struct.xnumber; RAISE NOTICE '[%]', lv_struct.xdate; RAISE NOTICE '[%]', lv_struct.xstring; END; $$;
It should return:
psql:verify_pg.SQL:263: NOTICE: [86] psql:verify_pg.SQL:263: NOTICE: [1944-04-25] psql:verify_pg.SQL:263: NOTICE: [Happy]
The program returns a structure with values converted into their appropriate data type.
- The second use-case checks with a query:
WITH get_struct AS (SELECT cast_strings(ARRAY['99','2015-06-14','Agent 99']) AS mystruct) SELECT (mystruct).xnumber , (mystruct).xdate , (mystruct).xstring FROM get_struct;
It should return:
xnumber | xdate | xstring ---------+------------+---------- 99 | 2015-06-14 | Agent 99 (1 row)
The query defines a call to the cast_strings function with a valid set of values and then displays the elements of the returned structure.
As always, I hope this helps those looking for how to solve this type of problem. Just a quick reminder that this was written and tested in PostgreSQL 14.
PL/pgSQL Date Function
This post provides an example of using PostgreSQL’s REGEXP_MATCH function, which works very much like the REGEXP_LIKE function in Oracle and a verify_date function that converts a string data type to date data type.
Here’s a basic function to show how to use a generic REGEXP_MATCH function:
1 2 3 4 5 6 7 8 9 10 11 | DO $$ DECLARE lv_date_in DATE := '2022-10-22'; BEGIN IF (REGEXP_MATCH('2022-10-02','^[0-9]{4,4}-[0-9]{2,2}-[0-9]{2,2}$') IS NOT NULL) THEN RAISE NOTICE '[%]', 'Truth'; END IF; END; $$; |
The following is a verify_date function, which takes a string with the ‘YYYY-MM-DD’ or ‘YY-MM-DD’ format and returns a BOOLEAN true or false value.
CREATE FUNCTION verify_date ( IN pv_date_in VARCHAR(10)) RETURNS BOOLEAN AS $$ DECLARE /* Local return variable. */ lv_retval BOOLEAN := FALSE; BEGIN /* Check for a YYYY-MM-DD or YYYY-MM-DD string. */ IF REGEXP_MATCH(pv_date_in,'^[0-9]{2,4}-[0-9]{2,2}-[0-9]{2,2}$') IS NOT NULL THEN /* Case statement checks for 28 or 29, 30, or 31 day month. */ CASE /* Valid 31 day month date value. */ WHEN (LENGTH(pv_date_in) = 10 AND SUBSTRING(pv_date_in,6,2) IN ('01','03','05','07','08','10','12') AND TO_NUMBER(SUBSTRING(pv_date_in,9,2),'99') BETWEEN 1 AND 31) OR (LENGTH(pv_date_in) = 8 AND SUBSTRING(pv_date_in,4,2) IN ('01','03','05','07','08','10','12') AND TO_NUMBER(SUBSTRING(pv_date_in,7,2),'99') BETWEEN 1 AND 31) THEN lv_retval := TRUE; /* Valid 30 day month date value. */ WHEN (LENGTH(pv_date_in) = 10 AND SUBSTRING(pv_date_in,6,2) IN ('04','06','09','11') AND TO_NUMBER(SUBSTRING(pv_date_in,9,2),'99') BETWEEN 1 AND 30) OR (LENGTH(pv_date_in) = 8 AND SUBSTRING(pv_date_in,4,2) IN ('04','06','09','11') AND TO_NUMBER(SUBSTRING(pv_date_in,7,2),'99') BETWEEN 1 AND 30) THEN lv_retval := TRUE; /* Valid 28 or 29 day month date value. */ WHEN (LENGTH(pv_date_in) = 10 AND SUBSTRING(pv_date_in,6,2) = '02') OR (LENGTH(pv_date_in) = 8 AND SUBSTRING(pv_date_in,4,2) = '02') THEN /* Verify 4-digit year. */ IF (LENGTH(pv_date_in) = 10 AND MOD(TO_NUMBER(SUBSTRING(pv_date_in,1,4),'99'),4) = 0 AND TO_NUMBER(SUBSTRING(pv_date_in,9,2),'99') BETWEEN 1 AND 29) OR (LENGTH(pv_date_in) = 8 AND MOD(TO_NUMBER(SUBSTRING(TO_CHAR(TO_DATE(pv_date_in,'YYYY-MM-DD'),'YYYY-MM-DD'),1,4),'99'),4) = 0 AND TO_NUMBER(SUBSTRING(pv_date_in,7,2),'99') BETWEEN 1 AND 29) THEN lv_retval := TRUE; ELSE /* Not a leap year. */ IF (LENGTH(pv_date_in) = 10 AND TO_NUMBER(SUBSTRING(pv_date_in,9,2),'99') BETWEEN 1 AND 28) OR (LENGTH(pv_date_in) = 8 AND TO_NUMBER(SUBSTRING(pv_date_in,7,2),'99') BETWEEN 1 AND 28)THEN lv_retval := TRUE; END IF; END IF; NULL; END CASE; END IF; /* Return date. */ RETURN lv_retval; END; $$ LANGUAGE plpgsql; |
The following four SQL test cases:
SELECT verify_date('2020-07-04') AS "verify_date('2020-07-04')"; SELECT verify_date('71-05-31') AS "verify_date('71-05-31')"; SELECT verify_date('2024-02-29') AS "verify_date('2024-02-29')"; SELECT verify_date('2019-04-31') AS "verify_date('2019-04-31')"; |
Return the following:
verify_date('2020-07-04')
---------------------------
t
(1 row)
verify_date('71-05-31')
-------------------------
t
(1 row)
verify_date('2024-02-29')
---------------------------
t
(1 row)
verify_date('2019-04-31')
---------------------------
f
(1 row) |
As always, I hope the example code fills somebody’s need.
PL/SQL List to Struct
Every now and then, I get questions from folks about how to tune in-memory elements of their PL/SQL programs. This blog post address one of those core issues that some PL/SQL programmers avoid.
Specifically, it addresses how to convert a list of values into a structure (in C/C++ its a struct, in Java its an ArrayList, and PL/SQL it’s a table of scalar or object types). Oracle lingo hides the similarity by calling either an Attribute Definition Type (ADT) or User-Defined Type (UDT). The difference in the Oracle space is that an ADT deals with a type defined in DBMS_STANDARD package, which is more or less like a primitive type in Java.
Oracle does this for two reasons:
- They handle lists of standard types in a difference C++ class than they do UDT types.
- They rigidly adhere to Interface Definition Language (IDL) principles.
The cast_strings function converts a list of strings into a record data structure. It lets the list of strings have either a densely or sparsely populated list of values, and it calls the verify_date function to identify a DATE data type and regular expressions to identify numbers and strings.
You need to build a UDT object type and lists of both ADT and UDT data types.
/* Create a table of strings. */ CREATE OR REPLACE TYPE tre AS TABLE OF VARCHAR2(20); / /* Create a structure of a date, number, and string. */ CREATE OR REPLACE TYPE struct IS OBJECT ( xdate DATE , xnumber NUMBER , xstring VARCHAR2(20)); / /* Create a table of tre type. */ CREATE OR REPLACE TYPE structs IS TABLE OF struct; / |
The cast_strings function is defined below:
CREATE OR REPLACE FUNCTION cast_strings ( pv_list TRE ) RETURN struct IS /* Declare a UDT and initialize an empty struct variable. */ lv_retval STRUCT := struct( xdate => NULL , xnumber => NULL , xstring => NULL); BEGIN /* Loop through list of values to find only the numbers. */ FOR i IN 1..pv_list.LAST LOOP /* Ensure that a sparsely populated list can't fail. */ IF pv_list.EXISTS(i) THEN /* Order if number evaluation before string evaluation. */ CASE WHEN lv_retval.xnumber IS NULL AND REGEXP_LIKE(pv_list(i),'^[[:digit:]]*$') THEN lv_retval.xnumber := pv_list(i); WHEN verify_date(pv_list(i)) THEN IF lv_retval.xdate IS NULL THEN lv_retval.xdate := pv_list(i); ELSE lv_retval.xdate := NULL; END IF; WHEN lv_retval.xstring IS NULL AND REGEXP_LIKE(pv_list(i),'^[[:alnum:]]*$') THEN lv_retval.xstring := pv_list(i); ELSE NULL; END CASE; END IF; END LOOP; /* Print the results. */ RETURN lv_retval; END; / |
There are three test cases for this function:
- The first use-case checks whether the input parameter is a sparsely or densely populated list:
DECLARE /* Declare an input variable of three or more elements. */ lv_list TRE := tre('Berlin','25','09-May-1945','45'); /* Declare a variable to hold the compound type values. */ lv_struct STRUCT; BEGIN /* Make the set sparsely populated. */ lv_list.DELETE(2); /* Test the cast_strings function. */ lv_struct := cast_strings(lv_list); /* Print the values of the compound variable. */ dbms_output.put_line(CHR(10)); dbms_output.put_line('xstring ['||lv_struct.xstring||']'); dbms_output.put_line('xdate ['||TO_CHAR(lv_struct.xdate,'DD-MON-YYYY')||']'); dbms_output.put_line('xnumber ['||lv_struct.xnumber||']'); END; /
It should return:
xstring [Berlin] xdate [09-MAY-1945] xnumber [45]
The program defines two numbers and deletes the first number, which is why it prints the second number.
- The second use-case checks with a list of only one element:
SELECT TO_CHAR(xdate,'DD-MON-YYYY') AS xdate , xnumber , xstring FROM TABLE(structs(cast_strings(tre('catch22','25','25-Nov-1945'))));
It should return:
XDATE XNUMBER XSTRING -------------------- ---------- -------------------- 25-NOV-1945 25 catch22
The program returns a structure with values converted into their appropriate data type.
- The third use-case checks with a list of two elements:
SELECT TO_CHAR(xdate,'DD-MON-YYYY') AS xdate , xnumber , xstring FROM TABLE(structs(cast_strings(tre('catch22','25','25-Nov-1945')) ,cast_strings(tre('31-APR-2017','1918','areodromes'))));
It should return:
XDATE XNUMBER XSTRING -------------------- ---------- -------------------- 25-NOV-1945 25 catch22 1918 areodromes
The program defines calls the cast_strings with a valid set of values and an invalid set of values. The invalid set of values contains a bad date in the set of values.
As always, I hope this helps those looking for how to solve this type of problem.
PL/SQL CASE Not Found
I was working on some test cases for my students and changing the behavior of a verify_date function that I wrote years ago to validate and returns valid dates when they’re passed as strings. The original program returned today’s date when the date was invalid.
The new function returns a BOOLEAN value of false by default and true when the string validates as a date. Unfortunately, I introduced a mistake that didn’t use to exist in Oracle 11g, which was the version when I wrote the original function.
The test cases in Oracle 21c raises the following error when an invalid date is passed to the CASE statement by the cast_strings function that calls the new verify_date function:
FROM TABLE(structs(cast_strings(tre('31-APR-2017','1917','dirk')))) * ERROR AT line 2: ORA-06592: CASE NOT found WHILE executing CASE statement ORA-06512: AT "C##STUDENT.VERIFY_DATE", line 30 ORA-06512: AT "C##STUDENT.CAST_STRINGS", line 18 |
As you can see, the test case uses ’31-APR-2017′ as an incorrect date to verify the use-case. The error occurred because the ELSE clause in the CASE statement wasn’t provided. Previously, the ELSE clause was optional and setting the lv_retval return variable to FALSE in the DECLARE block made it unnecessary.
The fixed code follows:
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 | CREATE OR REPLACE FUNCTION verify_date ( pv_date_in VARCHAR2) RETURN BOOLEAN IS /* Local variable to ensure case-insensitive comparison. */ lv_date_in VARCHAR2(11); /* Local return variable. */ lv_date BOOLEAN := FALSE; BEGIN /* Convert string input to uppercase month. */ lv_date_in := UPPER(pv_date_in); /* Check for a DD-MON-RR or DD-MON-YYYY string. */ IF REGEXP_LIKE(lv_date_in,'^[0-9]{2,2}-[ADFJMNOS][ACEOPU][BCGLNPRTVY]-([0-9]{2,2}|[0-9]{4,4})$') THEN /* Case statement checks for 28 or 29, 30, or 31 day month. */ CASE /* Valid 31 day month date value. */ WHEN SUBSTR(lv_date_in,4,3) IN ('JAN','MAR','MAY','JUL','AUG','OCT','DEC') AND TO_NUMBER(SUBSTR(pv_date_in,1,2)) BETWEEN 1 AND 31 THEN lv_date := TRUE; /* Valid 30 day month date value. */ WHEN SUBSTR(lv_date_in,4,3) IN ('APR','JUN','SEP','NOV') AND TO_NUMBER(SUBSTR(pv_date_in,1,2)) BETWEEN 1 AND 30 THEN lv_date := TRUE; /* Valid 28 or 29 day month date value. */ WHEN SUBSTR(lv_date_in,4,3) = 'FEB' THEN /* Verify 2-digit or 4-digit year. */ IF (LENGTH(pv_date_in) = 9 AND MOD(TO_NUMBER(SUBSTR(pv_date_in,8,2)) + 2000,4) = 0 OR LENGTH(pv_date_in) = 11 AND MOD(TO_NUMBER(SUBSTR(pv_date_in,8,4)),4) = 0) AND TO_NUMBER(SUBSTR(pv_date_in,1,2)) BETWEEN 1 AND 29 THEN lv_date := TRUE; ELSE /* Not a leap year. */ IF TO_NUMBER(SUBSTR(pv_date_in,1,2)) BETWEEN 1 AND 28 THEN lv_date := TRUE; END IF; END IF; ELSE NULL; END CASE; END IF; /* Return date. */ RETURN lv_date; EXCEPTION WHEN VALUE_ERROR THEN RETURN lv_date; END; / |
The new ELSE clause in on lines 31 and 32, and the converted function works. I also added a local lv_date_in variable to hold an uppercase version of an input string to: ensure a case-insensitive comparison of the month value, and avoid a having to pass the input as an IN OUT mode parameter. Typically, I leave off exception handlers because mistyping or copying for newer programmers becomes easier, but in this case I added an exception handler for strings that are larger than 11-characters.
As always, I hope this helps those looking for a solution to a coding problem.
Oracle DSN Security
Oracle disallows entry of a password value when configuring the ODBC’s Windows Data Source Name (DSN) configurations. As you can see from the dialog’s options:
So, I check the Oracle ODBC’s property list with the following PowerShell command:
Get-Item -Path Registry::HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\Oracle | Select-Object |
It returned:
Oracle Driver : C:\app\mclaughlinm\product\18.0.0\dbhomeXE\BIN\SQORA32.DLL
DisableRULEHint : T
Attributes : W
SQLTranslateErrors : F
LobPrefetchSize : 8192
AggregateSQLType : FLOAT
MaxTokenSize : 8192
FetchBufferSize : 64000
NumericSetting : NLS
ForceWCHAR : F
FailoverDelay : 10
FailoverRetryCount : 10
MetadataIdDefault : F
BindAsFLOAT : F
BindAsDATE : F
CloseCursor : F
EXECSchemaOpt :
EXECSyntax : F
Application Attributes : T
QueryTimeout : T
CacheBufferSize : 20
StatementCache : F
ResultSets : T
MaxLargeData : 0
UseOCIDescribeAny : F
Failover : T
Lobs : T
DisableMTS : T
DisableDPM : F
BatchAutocommitMode : IfAllSuccessful
Description : Oracle ODBC
ServerName : xe
Password :
UserID : c##student
DSN : Oracle |
Then, I used this PowerShell command to set the Password property:
Set-ItemProperty -Path Registry::HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\Oracle -Name "Password" -Value 'student' |
After setting the Password property’s value, I queried it with the following PowerShell command:
Get-ItemProperty -Path Registry::HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\Oracle | Select-Object -Property "Password" |
It returns:
Password : student |
After manually setting the Oracle ODBC DSN’s password value you can now connect without providing a password at runtime. It also means anybody who hacks the Windows environment can access the password through trivial PowerShell command.
I hope this alerts readers to a potential security risk when you use Oracle DSNs.
PL/pgSQL Coupled Loops
I love a challenge. A loyal Oracle PL/SQL developer said PL/pgSQL couldn’t support coupled loops and user-defined lists. Part true and part false. It’s true PL/pgSQL couldn’t support user-defined lists because it supports arrays. It’s false because PL/pgSQL supports an ARRAY_APPEND function that lets you manage arrays like Java’s ArrayList class.
Anyway, without further ado. You only need to create one data type because PL/pgSQL supports natural array syntax, like Java, C#, and other languages and doesn’t adhere rigidly to the Information Definition Language (IDL) standard that Oracle imposes. Oracle requires creating an Attribute Data Type (ADT) for the string collections, which you can avoid in PL/pgSQL.
You do need to create a record structure type, like:
/* Create a lyric object type. */ CREATE TYPE lyric AS ( day VARCHAR(8) , gift VARCHAR(24)); |
You can build a function to accept an array of strings and an array of record structures that returns a new array constructed from parts of the two input arrays. The function also compares and matches the two arrays before returning an array that combines strings for a songs lyrics. While the example uses the ever boring 12 Days of Christmas, I’d love another for examples. It just needs to use this type of repetitive structure. If you have one that you would like to share let me know.
The twelve_days function is:
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 | CREATE FUNCTION twelve_days ( IN pv_days VARCHAR(8)[] , IN pv_gifts LYRIC[] ) RETURNS VARCHAR[] AS $$ DECLARE /* Initialize the collection of lyrics. */ lv_retval VARCHAR(36)[114]; BEGIN /* Read forward through the days. */ FOR i IN 1..ARRAY_LENGTH(pv_days,1) LOOP lv_retval := ARRAY_APPEND(lv_retval,('On the ' || pv_days[i] || ' day of Christmas')::text); lv_retval := ARRAY_APPEND(lv_retval,('my true love sent to me:')::text); /* Read backward through the lyrics based on the ascending value of the day. */ FOR j IN REVERSE i..1 LOOP IF i = 1 THEN lv_retval := ARRAY_APPEND(lv_retval,('-'||'A'||' '|| pv_gifts[j].gift)::text); ELSIF j <= i THEN lv_retval := ARRAY_APPEND(lv_retval,('-'|| pv_gifts[j].day ||' '|| pv_gifts[j].gift )::text); END IF; END LOOP; /* A line break by verse. */ lv_retval := ARRAY_APPEND(lv_retval,' '::text); END LOOP; /* Return the song's lyrics. */ RETURN lv_retval; END; $$ LANGUAGE plpgsql; |
Then, you can test it with this query:
SELECT UNNEST(twelve_days(ARRAY['first','second','third','fourth' ,'fifth','sixth','seventh','eighth' ,'nineth','tenth','eleventh','twelfth'] ,ARRAY[('and a','Partridge in a pear tree')::lyric ,('Two','Turtle doves')::lyric ,('Three','French hens')::lyric ,('Four','Calling birds')::lyric ,('Five','Golden rings')::lyric ,('Six','Geese a laying')::lyric ,('Seven','Swans a swimming')::lyric ,('Eight','Maids a milking')::lyric ,('Nine','Ladies dancing')::lyric ,('Ten','Lords a leaping')::lyric ,('Eleven','Pipers piping')::lyric ,('Twelve','Drummers drumming')::lyric])) AS "12-Days of Christmas"; |
It prints:
12-Days of Christmas ---------------------------------- On the first day of Christmas my true love sent to me: -A Partridge in a pear tree On the second day of Christmas my true love sent to me: -Two Turtle doves -and a Partridge in a pear tree On the third day of Christmas my true love sent to me: -Three French hens -Two Turtle doves -and a Partridge in a pear tree ... Redacted for space ... On the twelfth day of Christmas my true love sent to me: -Twelve Drummers drumming -Eleven Pipers piping -Ten Lords a leaping -Nine Ladies dancing -Eight Maids a milking -Seven Swans a swimming -Six Geese a laying -Five Golden rings -Four Calling birds -Three French hens -Two Turtle doves -and a Partridge in a pear tree |
So, I believe that I met the challenge and hopefully provided a concrete example of some syntax that seems to be missing from most of the typical places.
PL/SQL List Function
Students wanted to see how to write PL/SQL functions that accept, process, and return lists of values. I thought it would be cool to also demonstrate coupling of loop behaviors and wrote the example using the 12-Days of Christmas lyrics.
The twelve_days function accepts two different collections. One is an Attribute Data Type (ADT) and the other a User-Defined Type (UDT). An ADT is based on a scalar data type, and a UDT is based on an object type. Object types are basically data structures, and they support both positional and named notation for variable assignments.
The twelve_days function returns a list of string, which is an ADT of the VARCHAR2 data type. Creating the ADT types is easy and a single step, like:
/* Create a days object type. */ CREATE OR REPLACE TYPE days IS TABLE OF VARCHAR2(8); / /* Create a string object type. */ CREATE OR REPLACE TYPE song IS TABLE OF VARCHAR2(36); / |
Creating the UDT is more complex and requires two steps. You need to create the UDT object type, or structure, and then the list based on the UDT object type, like:
/* Create a lyric object type. */ CREATE OR REPLACE TYPE lyric IS OBJECT ( DAY VARCHAR2(8) , gift VARCHAR2(24)); / /* Create a lyrics object type. */ CREATE OR REPLACE TYPE lyrics IS TABLE OF LYRIC; / |
Now, you can create the twelve_days function that uses these ADT and UDT types, like:
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 | CREATE OR REPLACE FUNCTION twelve_days ( pv_days DAYS , pv_gifts LYRICS ) RETURN song IS /* Initialize the collection of lyrics. */ lv_retval SONG := song(); /* Local procedure to add to the song. */ PROCEDURE ADD ( pv_input VARCHAR2 ) IS BEGIN lv_retval.EXTEND; lv_retval(lv_retval.COUNT) := pv_input; END ADD; BEGIN /* Read forward through the days. */ FOR i IN 1..pv_days.COUNT LOOP ADD('On the ' || pv_days(i) || ' day of Christmas'); ADD('my true love sent to me:'); /* Read backward through the lyrics based on the ascending value of the day. */ FOR j IN REVERSE 1..i LOOP IF i = 1 THEN ADD('-'||'A'||' '||pv_gifts(j).gift); ELSE ADD('-'||pv_gifts(j).DAY||' '||pv_gifts(j).gift); END IF; END LOOP; /* A line break by verse. */ ADD(CHR(13)); END LOOP; /* Return the song's lyrics. */ RETURN lv_retval; END; / |
You may notice the local add procedure on lines 10 thru 15. It lets you perform the two tasks required for populating an element in a SQL object type list in one line in the main body of the twelve_days function.
The add procedure first uses the EXTEND function to allocate space before assigning the input value to the newly allocated element in the list. Next, you can call the function inside the following SQL query:
SELECT column_value AS "12-Days of Christmas" FROM TABLE(twelve_days(days('first','second','third','fourth' ,'fifth','sixth','seventh','eighth' ,'nineth','tenth','eleventh','twelfth') ,lyrics(lyric(DAY => 'and a', gift => 'Partridge in a pear tree') ,lyric(DAY => 'Two', gift => 'Turtle doves') ,lyric(DAY => 'Three', gift => 'French hens') ,lyric(DAY => 'Four', gift => 'Calling birds') ,lyric(DAY => 'Five', gift => 'Golden rings' ) ,lyric(DAY => 'Six', gift => 'Geese a laying') ,lyric(DAY => 'Seven', gift => 'Swans a swimming') ,lyric(DAY => 'Eight', gift => 'Maids a milking') ,lyric(DAY => 'Nine', gift => 'Ladies dancing') ,lyric(DAY => 'Ten', gift => 'Lords a leaping') ,lyric(DAY => 'Eleven',gift => 'Pipers piping') ,lyric(DAY => 'Twelve',gift => 'Drummers drumming')))); |
It will print:
12-Days of Christmas ------------------------------------ On the first day of Christmas my true love sent to me: -A Partridge in a pear tree On the second day of Christmas my true love sent to me: -Two Turtle doves -and a Partridge in a pear tree On the third day of Christmas my true love sent to me: -Three French hens -Two Turtle doves -and a Partridge in a pear tree ... redacted for space ... On the twelfth day of Christmas my true love sent to me: -Twelve Drummers drumming -Eleven Pipers piping -Ten Lords a leaping -Nine Ladies dancing -Eight Maids a milking -Seven Swans a swimming -Six Geese a laying -Five Golden rings -Four Calling birds -Three French hens -Two Turtle doves -and a Partridge in a pear tree |
As always, I hope the example helps those looking for a solution to this type of problem.
MySQL Windows DSN
Almost a Ripley’s Believe It or Not. An prior data science student told me that his new IT department setup a Windows component that let him connect his Excel Spreadsheets to their production MySQL database without a password. Intrigued, I asked if it was a MySQL Connector/ODBC Data Source Configuration, or DSN (Data Source Name)?
He wasn’t sure, so I asked him to connect to PowerShell and run the following command:
Get-Item -Path Registry::HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\MySQL |
It returned something like this (substituting output from one of my test systems):
Hive: HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI
Name Property
---- --------
MySQL Driver : C:\Program Files\MySQL\Connector ODBC 8.0\myodbc8w.dll
DESCRIPTION : MySQL ODBC Connector
SERVER : localhost
UID : student
PWD : student
DATABASE : studentdb
PORT : 3306 |
The student was stunned and concerned he was compromising his employer’s system security. I suggested he share the information with his IT department so they could provide a different approach for his access to the production database. His IT department immediately agreed. Unfortunately, he’s bummed he can’t simply access the data through Excel.
I told him they were welcome to use the MySQL Connect Dialog PowerShell solution that I wrote. It creates a minimal MySQL DSN and requires a manual password entry through the PowerShell Dialog box. I also suggested that they look into the PowerShell Excel Module.
I also suggested they develop a query only copy of the production database, or shift access to a data warehouse. Needless to say, it wasn’t a large corporation.
As always, I hope this helps others.
Subscribe via RSS