yhat / pandasql

sqldf for pandas
MIT License
1.31k stars 184 forks source link

Support of 'where exists' operations #73

Open MartinPetrov8 opened 5 years ago

MartinPetrov8 commented 5 years ago

Hi, i have a query i am trying to execute in pandasql but my jupyter environment freezes so i am assuming this is not supported. Query trying to execute:

q1=""" SELECT DATE(Timestamp) AS date, COUNT(DISTINCT User) AS user_count FROM df pc WHERE EXISTS ( SELECT * FROM df pc1 WHERE pc1.User = pc.User AND DATE(pc1.Timestamp) < DATE(pc.Timestamp) ) GROUP BY pc.Timestamp """

aijanai commented 4 years ago

where exists launches subqueries for each row, are you sure it isn't just computing a goatload of runs?