I was having flashbacks to doing fuzzy matches in sql, then realized that the number of emojis isn't too crazy. Could probably cache the common searches people are doing too.
Oh yeah, I guess fuzzy search is the wrong term to use as it implies the SQL tools. What would really work is just a list of alternative names that the emote goes by attached to it's record.
Or if possible a hashed table where each alternative name returns a pointer to the correct name.
Oh yeah, I guess fuzzy search is the wrong term to use as it implies the SQL tools
Nah, that's not what I meant. I've just had to do fuzzy search before in sql, and the issue is that fuzzy matching doesn't scale well. It's nightmarish how long it takes if you don't do proper filtering ahead of time (and have a table with 10s of millions of rows).
My brief foray into SQL and data engineering through ArcGIS has me absolutely amazed at how seamless so much of the internet is. Even doing a spatial query of a small town can take several minutes for a dataset with say 30,000 records. I tried some of my usual geospatial join and query techniques on a dataset of 175 million and my timeframe quickly approached the heat death of the universe.
I was having flashbacks to doing fuzzy matches in sql, then realized that the number of emojis isn't too crazy. Could probably cache the common searches people are doing too.
Oh yeah, I guess fuzzy search is the wrong term to use as it implies the SQL tools. What would really work is just a list of alternative names that the emote goes by attached to it's record.
Or if possible a hashed table where each alternative name returns a pointer to the correct name.
Nah, that's not what I meant. I've just had to do fuzzy search before in sql, and the issue is that fuzzy matching doesn't scale well. It's nightmarish how long it takes if you don't do proper filtering ahead of time (and have a table with 10s of millions of rows).
My brief foray into SQL and data engineering through ArcGIS has me absolutely amazed at how seamless so much of the internet is. Even doing a spatial query of a small town can take several minutes for a dataset with say 30,000 records. I tried some of my usual geospatial join and query techniques on a dataset of 175 million and my timeframe quickly approached the heat death of the universe.