Change Details

I am doing something like this: research_prod@db1108.eqiad.wmnet(log)>select -> concat(wiki, ': ', event_title) as title, -> event_token, -> event_context -> from -> UniversalLanguageSelector_17799034 -> where -> event_token = 'sdfs98df7s9d8f9s'; +------------------+------------------+---------------+ | title | event_token | event_context | +------------------+------------------+---------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | Armenian | | enwiki: Moscow | sdfs98df7s9d8f9s | հա | | enwiki: Moscow | sdfs98df7s9d8f9s | hayeren | +------------------+------------------+---------------+ I'd like to get something like this: +-------------------+------------------+-------------------------------+ | title | event_token | event_context | +-------------------+------------------+-------------------------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | ["Armenian", "հա", "hayeren"] | +-------------------+------------------+-------------------------------+ The second is preferable because I need to do it for a lot of tokens, and to analyze the data from event_context for each of them. Using GROUP_CONCAT is not great, because this is user input, and there is no safe way to concatenate it. JSON_ARRAYAGG should be safer.

I am doing something like this (this is not actual data; otherwise it would be private. I changed the values): research_prod@db1108.eqiad.wmnet(log)>select -> concat(wiki, ': ', event_title) as title, -> event_token, -> event_context -> from -> UniversalLanguageSelector_17799034 -> where -> event_token = 'sdfs98df7s9d8f9s'; +------------------+------------------+---------------+ | title | event_token | event_context | +------------------+------------------+---------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | Armenian | | enwiki: Moscow | sdfs98df7s9d8f9s | հա | | enwiki: Moscow | sdfs98df7s9d8f9s | hayeren | +------------------+------------------+---------------+ I'd like to get something like this: +-------------------+------------------+-------------------------------+ | title | event_token | event_context | +-------------------+------------------+-------------------------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | ["Armenian", "հա", "hayeren"] | +-------------------+------------------+-------------------------------+ The second is preferable because I need to do it for a lot of tokens, and to analyze the data from event_context for each of them. Using GROUP_CONCAT is not great, because this is user input, and there is no safe way to concatenate it. JSON_ARRAYAGG should be safer.

I am doing something like this (this is not actual data; otherwise it would be private. I changed the values): research_prod@db1108.eqiad.wmnet(log)>select -> concat(wiki, ': ', event_title) as title, -> event_token, -> event_context -> from -> UniversalLanguageSelector_17799034 -> where -> event_token = 'sdfs98df7s9d8f9s'; +------------------+------------------+---------------+ | title | event_token | event_context | +------------------+------------------+---------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | Armenian | | enwiki: Moscow | sdfs98df7s9d8f9s | հա | | enwiki: Moscow | sdfs98df7s9d8f9s | hayeren | +------------------+------------------+---------------+ I'd like to get something like this: +-------------------+------------------+-------------------------------+ | title | event_token | event_context | +-------------------+------------------+-------------------------------+ | enwiki: Moscow | sdfs98df7s9d8f9s | ["Armenian", "հա", "hayeren"] | +-------------------+------------------+-------------------------------+ The second is preferable because I need to do it for a lot of tokens, and to analyze the data from event_context for each of them. Using GROUP_CONCAT is not great, because this is user input, and there is no safe way to concatenate it. JSON_ARRAYAGG should be safer.