I have a following string (well, UTM actually):
bla-ble-blo-bli-this-bad-blau-bleu
bla-ble-blo-bli-this_good-blau-bleu
5th position in UTM is called "Subject", and currently since I am using SQL to do any sort of transformations, I am using a very simple set of rules and some look-up table as I am splitting string on "-", but as you see, it fails on "this-bad" as it has a "-", but it works as expected on "this_good".
Is there a regex to get 5th occurence of this "Subject" part? I tried following regex, but I really don't know how to set it find only 5th occurence. The regex in question: \w+(?=-[^-]*$).
Thank you!
Remove the tail and head, leaving the 5th part:
select regexp_replace(regexp_replace('bla-ble-blo-bli-this-bad-blau-bleu', '(-[^-]+){2}$', ''), '([^-]+-){4}', '')
See live demo.
Related
I need to retrieve the bolded section of the below string . This value is in a column within my Postgres database table.
SEALS_LME_TRADES_MBL_20220919_00212.csv
I tried to utilize the functions; substring, reverse, strpos but they all have limitations. It seems like regex is the best option, however I was not able to do it.
Essentially I need to substring from beginning till the second last '_'. I do not want the date and sequence number along with the file extension at the end.
The closes regex I managed to get is: ^(([^]*){4})
https://regex101.com/
This look a little wonky but how about this?
select substring ('SEALS_LME_TRADES_MBL_20220919_00212.csv', '^(.+)_[^_]+_[^_]+')
Translation
^ from the beginning
(.+) any characters (capture and return this value), followed by
_ an underscore, followed by
[^_]+ one or more non-underscores, followed by
_ an underscore, followed by
[^_]+ one or more non-underscores
Regex greediness will cause any incidental underscores to be captured in the initial string.
Technically speaking the last portion (one or more non-underscores) can probably be omitted.
I am trying to extract a item_subtype field from an URL.
This regex works fine in the to get the first item item_type
SELECT REGEXP_EXTRACT('info?item_type=icecream&item_subtype=chocolate/cookies%20cream,vanilla&page=1', r'item_type=(\w+)')
but what is the correct regex to get everything starting from 'chocolate' all the way to before the '&page1'
I have tried this, but can't seem to get it to work to go further
SELECT REGEXP_EXTRACT('info?item_type=icecream&item_subtype=chocolate/cookies%20cream,vanilla&page=1', r'item_subtype=(\w+[^Z])')
basically, I want to extract 'chocolate/cookies%20cream,vanilla'
In your case, \w+ only matches one or more letters, digits or underscores. Your expected values may contain other characters, too.
You may use
SELECT REGEXP_EXTRACT('info?item_type=icecream&item_subtype=chocolate/cookies%20cream,vanilla&page=1', r'item_subtype=([^&]+)')
See the regex demo.
Notes:
item_subtype= - this string is matched as a literal char sequence
([^&]+) - a Capturing group 1 that matches and captures one or more chars other than & into a separate memory buffer that is returned by REGEXP_EXTRACT function.
In Bigquery, I am trying to find a way to extract particular segments of a string based on how many dashes come before it. The number of total dashes in the string will always be the same. For example, I could be looking for the string after the second dash and before the third dash in the following string:
abc-defgh-hij-kl-mnop
Currently, I am using the following regex to extract, which counts the dashes from the back:
([^-]+)(?:-[^-]+){2}$
The problem is that if there is nothing in between the dashes, the regex doesn't work. For example, something like this returns null:
abc-defgh-hij--mnop
Is there a way to use regex to extract a string after a certain number of dashes and cut it off before the subsequent dash?
Thank you!
Below is for BigQuery Standrd SQL
The simplest way in your case is to use SPLIT and OFFSET as in below example
SELECT SPLIT(str, '-')[OFFSET(3)]
above will return empty string for abc-defgh-hij--mnop
to prevent error in case of calling non-existing element - better to use SAFE_OFFSET
SELECT SPLIT(str, '-')[SAFE_OFFSET(3)]
Have strings containing 'q_' which I want to extract everything that comes after it. Some rows contain occurrence of q_ which I want everything that occurs after it. Example values in the column are:
prod-q_cat_trait_cat_social_issue
_prod-q_body_modification_graffiti
event_tickets
dappled_grey
_prod-q_cat_tech_support
What is wrong with my regular expression as I'm trying to remove the trailing '_' after q.
REGEXP_EXTRACT(queue_id, '[^q_]+$')
Is just returning
issue
I've also tried the split method:
SPLIT(queue_id, 'q_')[OFFSET(2)]
But this returns
Array index 2 is out of bounds (overflow)
Any suggestions. Thanks! (I am using Google Cloud SQL)
Using a capturing group, you may extract all after the first q_ with:
REGEXP_EXTRACT(queue_id, 'q_(.*)')
You may extract all after the last q_ with:
REGEXP_EXTRACT(queue_id, '.*q_(.*)')
See the regex demo #1 and regex demo #2.
Here, q_ finds the first occurrence of q_ and (.*) grabs the rest of the line into Group 1, and this is the value returned by REGEXP_EXTRACT. .* matches any 0+ chars other than line break chars as many as possible, that is why the second regex will start capturing the rest of the line after the last occurrence of q_.
Google Cloud SQL uses MySQL. I think the simplest method is substring_index():
select substring_index(queue_id, '-q_', -1)
Can you try this : q_([^q_]+)$? You'll have what you want in the first group.
Edit: this one match all the cases > (?(?<=-q_).*|^((?!-q_).)*$)
I have a column in Openrefine, which I would like to add a character string in each of its rows, based on the position in the string.
For example:
I have an 8th character number string: 85285296 and would like to add "-" at the fourth place: "8528-5296".
Anyone can help me find the specific function in OpenRefine?
Thanks
Tzipy
The simplest approach is to just use the expression language's built-in string indexing and concatenation:
value[0,4]+'-'+value[4,8]
or more generally, if you don't know that your value is exactly 8 characters long:
value[0,4]+'-'+value[4,999]
Possible solution (not sure if it's the most straightforward):
value.replace(/(\d{4})(.+)/, "$1-$2")
This means : if $1 represents the content of the first parenthesis/group in the regular expression before and $2 the content of the second one, replaces each value in the column with $1-$2.
Some other options:
value.splitByLengths(4,4).join("-")
value.match(/(\d{4})(\d{4})/).join("-")
value.substring(0,4)+"-"+value.substring(4,8)
I think 'splitByLengths' is the neatest, but I might use 'match' instead because it fails with an error if your starting string isn't 8 digits - which means you don't accidentally process data that doesn't conform to your assumption of what data is in the column - but you could use a facet/filter to check this with any of the others