General problem: I want to do a left join on two huge tables, but there is no matching key - I want to join the left table with the "nearest" row in the right table. In this example I want to join using a timestamp, but I think one would have a similar (though more complex) problem with geographical coordinates. To simplify a lot, I have one transaction table with a primary key id, timestamp, currency and amount - and one currency rate table with id, timestamp, currency and rate. Those tables are independent from each other, and they contain a lot of rows; assume that for some periods there are many transactions between each update of the currency rate, while in other periods there are many updates of the currency rates for each transaction. I want to join them so that there is one row for each transaction, with the closest matching currency rate. I have indexes on the timestamps. So, here is a brief example of what I have and what I need: transaction: +---------------------+--------+ | timestamp | amount | +---------------------+--------+ | 2017-07-31 00:07:05 | 173.17 | | 2017-08-02 15:29:11 | 136.57 | | 2017-08-05 03:45:24 | 81.27 | | 2017-08-05 03:46:47 | 48.1 | | 2017-08-05 03:47:38 | 35.21 | +---------------------+--------+ rate: +---------------------+-------------------+ | timestamp | rate | +---------------------+-------------------+ | 2017-07-31 00:04:49 | 9.2923 | | 2017-07-31 01:37:59 | 9.313423370522607 | | 2017-08-01 08:07:59 | 9.325 | | 2017-08-01 16:52:23 | 9.3542 | | 2017-08-01 21:07:09 | 9.357076262185192 | | 2017-08-02 15:07:33 | 9.34936993421895 | | 2017-08-02 17:48:45 | 9.357217848393876 | | 2017-08-04 04:33:51 | 9.38690807898736 | | 2017-08-04 08:13:01 | 9.383765889641701 | | 2017-08-06 03:45:03 | 9.118193727124817 | | 2017-08-06 04:15:01 | 9.353042966450854 | | 2017-08-06 05:23:29 | 9.353042966450854 | +---------------------+-------------------+ Want something like this: +---------------------+--------+--------+ | timestamp | amount | rate | +---------------------+--------+--------+ | 2017-07-31 00:07:05 | 173.17 | 9.2923 | | 2017-08-02 15:29:11 | 136.57 | 9.3493 | | 2017-08-05 03:45:24 | 81.27 | 9.3869 | | 2017-08-05 03:46:47 | 48.1 | 9.3869 | | 2017-08-05 03:47:38 | 35.21 | 9.3869 | +---------------------+--------+--------+ (The production data is much more dense than this)