In our first weblog, we used an everyday expression to switch the quotes in genres. Afterward, we have been in a position to UNNEST() the JSON object. We’ll be working with the identical information set on this weblog
In our information:
Embedded content material: https://gist.github.com/nfarah86/ef1cc9da88e56226c4c46fd0e3c8e16e
there’s a JSON string that’s referred to as spoken_languages, and it’s formatted equally to genres:
[ { "spoken_languages": "[{'iso_639_1': 'fr', 'name': 'Français'}]" }]
Assuming all the things is constant, we will simply write the SQL assertion just like what we wrote for genres— proper?
Improper ⛔️
We truly get a parsing error:
json parse error on line 0 close to `x9akai”}, {“iso_’: unknown escape 120 in string
So, we type of know what the wrongdoer is right here, however we don’t know if there are going to be extra parsing errors. Let’s go forward and debug this, so we will get this SQL assertion working!
We will truly use REGEXP_LIKE() to see what precisely is inflicting the error:
SELECT
spoken_languages
FROM
commons.TwtichMovies t
WHERE REGEXP_LIKE(t.spoken_languages,'x9akai')
This can be a pattern of what we get again:
That is nice— it seems to be just like the are literally inflicting some points. We will use REGEXP_REPLACE() to switch these slashes with an empty string:
SELECT (REGEXP_REPLACE('Lietuvix9akai', '',''));
Now— to place all of it collectively: How can we UNNEST() spoken_languages and repair the two points at hand (the string format and the double slashes)?
Trace:
SELCT REGEXP_REPLACE(REGEXP_REPLACE('[{''iso_639_1'': ''lt'', ''name'':''Lietuvix9akai''}]', '''', '"'), '', '');
I’m positive you possibly can take it to the end line from right here, however simply in case, you possibly can watch this youtube hyperlink down under to catch the total replay!
Embedded content material: https://youtu.be/8aHgJrQjT4U
TLDR: yow will discover all of the sources you want within the developer nook.
