Postgres SQL - different results from LIKE query using OR vs || - postgresql

I have a table with an integer column. It has 12 records numbered 1000 to 1012. Remember, these are ints.
This query returns, as expected, 12 results:
select count(*) from proposals where qd_number::text like '%10%'
as does this:
SELECT COUNT(*) FROM "proposals" WHERE (lower(first_name) LIKE '%10%' OR qd_number::text LIKE '%10%' )
but this query returns 2 records:
SELECT COUNT(*) FROM "proposals" WHERE (lower(first_name) || ' ' || qd_number::text LIKE '%10%' )
which implies using || in concatenated where expressions is not equivalent to using OR. Is that correct or am I missing something else here?

You probably have nulls in first_name. For these records (lower(first_name) || ' ' || qd_number::text results in null, so you don't find the numbers any longer.

using || in concatenated where expressions is not equivalent to using ORIs that correct or am I missing something else here?
That is correct.
|| is the string concatenation operator in SQL, not the OR operator.

Related

Postgresql - How to make NOT LIKE + SELECT statement, comparing two columns between tables

Hit a bit of a brickwall here, I haven't found anything that seems to work.
SELECT *
FROM glsltransaction gls
INNER JOIN glhistory h ON gls.sltrxid = h.schedxrefid
INNER JOIN apvendor ap ON ap.vendorid = gls.acctid
INNER JOIN glchartofaccounts coa USING (acctdeptid)
WHERE h.description <> gls.description
AND h.description not like || ap.name || '%'
AND gls.description <> ''
This line here AND h.description not like || ap.name || '%' is what i'm having issues with.
I get the following error when trying to run that statement above:
'Error occurred in running query from editor : ERROR: operator does not exist: || text Hint: No operator matches the given name and argument type. You might need to add an explicit type cast. Position: 563'
I'm effectively wanting it to function like a.column not like b.column%
Any help will be greatly appreciated, thanks!

Postgresql SQLSTATE[42P18]: Indeterminate datatype with PDO and CONCAT

I'm having issues with CONCAT() when used on a WHERE, in PDO.
The code:
<?php
require_once('config.php');
$fdate = '01/01/2010';
$tdate = '31/12/2030';
$identification = '';
$count = "SELECT count(*) as total FROM ( select time_id from doc_sent WHERE date >= :fdate AND date <= :tdate AND identification LIKE concat('%',:identification,'%') ) x;";
//$count = "SELECT count(*) as total FROM ( select time_id from doc_sent WHERE date >= :fdate AND date <= :tdate ) x;";
$stmt_count_row_main_table = $pdo->prepare($count);
$stmt_count_row_main_table->execute(['fdate' => $fdate, 'tdate' => $tdate, 'identification' => $identification]);
//$stmt_count_row_main_table->execute(['fdate' => $fdate, 'tdate' => $tdate]);
$count_row_main_table = $stmt_count_row_main_table->fetch();
print_r( $count_row_main_table);
?>
The code works when the 'identification' part is commented.
When I'm trying to use CONCAT(), it doesn't.
I tried many "version" of CONCAT() (and read many other questions, like this one: How do I create a PDO parameterized query with a LIKE statement? ) but I am always referring to the main documentation:
https://www.postgresql.org/docs/9.1/static/functions-string.html
Which say:
concat('abcde', 2, NULL, 22) --> abcde222
The FULL error when I use CONCAT() is:
PHP Fatal error: Uncaught PDOException: SQLSTATE[42P18]: Indeterminate datatype: 7 ERROR: could not determine data type of parameter $3 in /var/www/pdo-reporter/show.php:17\nStack trace:\n#0 /var/www/pdo-reporter/show.php(17): PDOStatement->execute(Array)\n#1 {main}\n thrown in /var/www/pdo-reporter/show.php on line 17
What's wrong with my code?
CONCAT is a function that takes a VARIADIC argument list, which means that internally postgres will convert them into an array of the same type.
postgres=# \df concat
List of functions
Schema | Name | Result data type | Argument data types | Type
------------+--------+------------------+---------------------+------
pg_catalog | concat | text | VARIADIC "any" | func
When trying to resolve the input type to a single type, the SQL parser fails. It can be reproduced in this simpler form:
postgres=# PREPARE p AS select concat('A', $1);
ERROR: could not determine data type of parameter $1
The parser can't figure out the datatype of $1 so it errs on the side of caution.
One easy solution is to cast the parameter as text:
postgres=# PREPARE p AS select concat($1::text);
PREPARE
or with the CAST operator:
postgres=# PREPARE p AS select concat(cast($1 as text));
PREPARE
I haven't tested with PDO but presumably it would work (given how it deals with parameters to produce prepared statements) to change the query to:
"...identification LIKE '%' || :identification || '::text%'..."
or use the '||' operator instead of concat in the query:
identification LIKE '%' || :identification || '%'
EDIT: BTW if you want to find that a parameter :X is a substring of identification, this clause is more secure: strpos(identification, :X) > 0, because :X may contain '%' or '_' without causing any side-effect in the match, contrary to what happens with LIKE.

return one string from colums data

I've result set from query select * from personal."phoneNumbers" like this
prefix
pref |number
-----|--------
"12 "|"4589524"
"077"|"7090701"
"050"|"2561024"
But I want to return data like
(12) 4589524;(077) 7090701; (050) 2561024
How to do this with postgresql ?
You can use the string operators and functions to construct a single phone number in the format that you want:
'(' || btrim(pref) || ') ' || number
This, obviously, yields a string for each record that you process. You can then use the aggregation function string_agg() to string (no pun intended) the extended phone numbers from all the records together into one, with the appropriate separator between phone numbers:
SELECT string_agg('(' || btrim(pref) || ') ' || number, '; ') AS pref_number
FROM personal."phoneNumbers"

How to translate the PostgreSQL array_agg function to SQLite?

This query works in PostgreSQL:
Select ot.MCode,array_to_string(array_agg(tk1.TName || ',' || ot.TTime), ' - ') as oujyu_name_list
From TR_A ot
inner join MS_B tk1 on ot.Code = tk1.Code
Where ot.Code in (Select Code From TR_C )
Group byot.MCode
but it does not work in SQLite, because SQLite does not have the array_agg() function. How can this query be converted to SQLite?
For this query, you can use group_concat, which directly returns a string:
SELECT ..., group_concat(tk1.TName || ',' || ot.TTime, ' - ')
FROM ...
SQLite now has the JSON1 extension (which ships in the default distribution) that can group and create arrays of JSON objects. For example,
select
ot.MCode,
json_group_array(json_object('tname', tk1.TName, 'ttime', ot.TTime)) as oujyu_name_list
from TR_A as ot
inner join MS_B as tk1
on (ot.Code = tk1.Code)
where ot.Code in (select code from TR_C)
group by ot.MCode;
The second column will be formatted as a JSON array e.g. [{"tname":...,"ttime":...},...].

Postgresql SELECT if string contains

So I have a in my Postgresql:
TAG_TABLE
==========================
id tag_name
--------------------------
1 aaa
2 bbb
3 ccc
To simplify my problem,
What I want to do is SELECT 'id' from TAG_TABLE when a string "aaaaaaaa" contains the 'tag_name'.
So ideally, it should only return "1", which is the ID for tag name 'aaa'
This is what I am doing so far:
SELECT id FROM TAG_TABLE WHERE 'aaaaaaaaaaa' LIKE '%tag_name%'
But obviously, this does not work, since the postgres thinks that '%tag_name%' means a pattern containing the substring 'tag_name' instead of the actual data value under that column.
How do I pass the tag_name to the pattern??
You should use tag_name outside of quotes; then it's interpreted as a field of the record. Concatenate using '||' with the literal percent signs:
SELECT id FROM TAG_TABLE WHERE 'aaaaaaaa' LIKE '%' || tag_name || '%';
And remember that LIKE is case-sensitive. If you need a case-insensitive comparison, you could do this:
SELECT id FROM TAG_TABLE WHERE 'aaaaaaaa' LIKE '%' || LOWER(tag_name) || '%';
A proper way to search for a substring is to use position function instead of like expression, which requires escaping %, _ and an escape character (\ by default):
SELECT id FROM TAG_TABLE WHERE position(tag_name in 'aaaaaaaaaaa')>0;
I personally prefer the simpler syntax of the ~ operator.
SELECT id FROM TAG_TABLE WHERE 'aaaaaaaa' ~ tag_name;
Worth reading through Difference between LIKE and ~ in Postgres to understand the difference.
`
In addition to the solution with 'aaaaaaaa' LIKE '%' || tag_name || '%' there
are position (reversed order of args) and strpos.
SELECT id FROM TAG_TABLE WHERE strpos('aaaaaaaa', tag_name) > 0
Besides what is more efficient (LIKE looks less efficient, but an index might change things), there is a very minor issue with LIKE: tag_name of course should not contain % and especially _ (single char wildcard), to give no false positives.
SELECT id FROM TAG_TABLE WHERE 'aaaaaaaa' LIKE '%' || "tag_name" || '%';
tag_name should be in quotation otherwise it will give error as tag_name doest not exist