I want to select all rows that contains an string (for example abc)
example:
abc
abcd
0abc
0abcd
I want all the above to be returned.
My first approach was:
SELECT * FROM notes WHERE notes MATCH '^abc*'
but it returns 0 results.
Second was:
SELECT * FROM notes WHERE notes MATCH '*abc*'
but it returns an error (I belive that asterix can't be used as the first character).
How can I do this?
Thanks in advance
The issue is your usage of ^: your string does not start with abc, so it's not a match. According to SQLite documentation:
MATCH '^one' -- first token in any column must be "one"
Related
My data is string like:
'湯姆 is a boy.'
or '梅isagirl.'
or '約翰,is,a,boy.'.
And I want to split the string and only choose the Chinese name.
In R, I can use the command
tmp=strsplit(string,[A-z% ])
unlist(lapply(tmp,function(x)x[1]))
And then getting the Chinese name I want.
But in PostgreSQL
select regexp_split_to_array(string,'[A-z% ]') from db.table
I get a array like {'湯姆','','',''},{'梅','','',''},...
And I don't know how to choose the item in the array.
I try to use the command
select regexp_split_to_array(string,'[A-z% ]')[1] from db.table
and I get an error.
I don't think that regexp_split_to_array is the appropriate function for what you are trying to do here. Instead, use regexp_replace to selectively remove all ASCII characters:
SELECT string, regexp_replace(string, '[[:ascii:]~:;,"]+', '', 'g') AS name
FROM yourTable;
Demo
Note that you might have to adjust the set of characters to be removed, depending on what other non Chinese characters you expect to have in the string column. This answer gives you a general suggestion for how you might proceed here.
A particular column in my table is marked as a String type, however, more than likely, it will have an integer value mixed in with the string.
Here are some example rows:
"hello1"
"keys"
"Thom27"
"3for5"
I would like to be able to select the rows that have numbers in the string. Thank you.
You can use GLOB for that:
SELECT *
FROM tblNames
WHERE Name GLOB '*[0-9]*'
GLOB is the best answer (as mentioned above). GLOB supports REGEX like matching, which is more efficient than individual matches with LIKE.
Suppose I have such data:
Lena
Lera
Elena
Mark
Allen
Paul
When user enters 'le' I need to return
Find words by first letters
After that should follow all another words that contains 'le' chars (full text search)
So it should return:
Lena
Lera
Alen
Elena
Something like that (this example does not return anything):
SELECT name FROM table WHERE name LIKE 'le%' AND like '%le%' ORDER BY name ASC
Thank you.
After it turns out that UNION mixes up the order of the sub selects, we'll try something else. You could try it with a custom ORDER BY. First order by whether it starts with your search term, then by name.
SELECT name
FROM table
WHERE UPPER(name) LIKE UPPER('%le%')
ORDER BY (CASE WHEN UPPER(name) LIKE UPPER('le%') THEN 1 ELSE 2 END),
name ASC
Comparing the strings with UPPER helps ignoring the case. But UPPER only support ASCII apparently, read this article for more information on the topic. It also contains other ways to ignore case when comparing strings.
I have a column in a teradata table with string values like "page1-->page2-->page1-->page3-->page1--page2-->..."
I want to search for a specific page and get the number of occurrence of the page in the string. I couldn't find any function that gives this result.
There's no builtin function, but there's a common solution:
Remove all occurences of the substring from the string and compare the length before/after:
(Char_Length(string) - Char_Length(OReplace(string, searchstr))) / Char_Length(searchstr)
Edit:
For a wildcard search you can utilize REGEXP_REPLACE:
Char_Length(RegExp_Replace(RegExp_Replace(s, 'page1(.+?)page3', '#',1,0), '[^#]','',1,0))
For `#' use a character which is known not to be in your input string.
I Have multiple records in table like below. Each record holds mutiple entries separated by #.
record1 - 123.45.56:ABCD:789:E # 1011.1213.1415:FGHI:1617:J #
record2 - 123.45.56:ABCD:1617:E # 1011.1213.1415:FGHI:12345:J #
I need to pass an argument to a different project/service which builds an sql query and send the output to me.
Now if I send an argument like below, it gives me wrong output
123.45.56:*:1617
This recognizes both record1 and record 2 as proper output because of wildcard char. But as per my requirement only record2 is proper as record1 has 123.45.56 in one entry and 1617 in a different entry.
Is there a way to construct an expression that says the like condition to ignore such invalid entries.
Please note that I cant change the query as I am not constructing it. The only way for me is to tweak the expression that I can send as argument.
You need to restrict the pattern you match to be specic enough such that it only matches the first record and not the second one.
You can try:
SELECT *
FROM yourTable
WHERE col LIKE '123.45.56:' AND col LIKE '1617:J #'