All right, let's talk about the pricing concepts. So there are three major categories of pricing. So as we talked about you have, your storage amount, which is how much data that you're storing. Also, included in there is, if you're streaming data ends if you're not uploading it into a like a batch, which is free sub, like uploading a five GB CSV file at once, that's batch. Now, if you're streaming say you have like an API that says, as soon as this event happens, I want you to write this individual record of these 10 records into the BigQuery as it happens. You will be actually charged for streaming inserts, but you can stream data into BigQuery in your old time. You see at the bottom, you get an automatic discount for all data. So one of the cool things is, our view right now it's two cents per GB as of right now since October, 2017. Two cents per GB per month to store data. But, if it has not been changed for a consecutive 90 days, you go into what's called the long-term storage, and the amount that it costs to store that data per month is cut in half. So, it's instead of two cents per GB, it's one cent per GB. Now, for processing, this is where you're going to incur the majority of your charges. So, if you're doing optimization, you want to make sure that you write amazingly fast and beautiful queries first, before you talk about, "Oh, maybe do we need all of these tables, let's cut out a few, " because your going to get the biggest gains by optimizing your queries. So, for the majority of everyone, you're going to using the on-demand, or pay-as-you-go plan. So bytes that you process, that's going to be what you're going to be charged. If you're an extremely large organization, there's a flat-rate plan. For those folks who just want to have a set amount of cost for BigQuery normally like a 10 or $20,000 per month. But, again all those pricing plans that are available online with the latest pricing details. So, we're largely going to be focused on on-demand pricing, and in the lab, we're going to be focusing on taking the amount of bytes that you're going to process, plugging in into a pricing calculator that's available online, and then seeing how much it actually is cost to process. The good news again, there is the free component of this, to get you ramped up. So you have 1 TB, or you know just over a 1,000 GB of data per month, for processing, and even in storage you have 10 GB of free storage. So with our IRS datasets that are in the megabytes, like not even a GB, you get 10 GB free to start you out for storage, and a TB of query processing, which is great. You have to opt-in to one high-compute queries. So, high-compute queries, instead of using a lot of storage, you're using a lot of computationally intensive things like, user defined, functions, potentially, or other things that just require a ton of mathematical operation. And, last but not the least, as we mentioned exporting, loading, data, queries on metadata. So, this is an interesting point. So BigQuery by nature, as you saw in the preview, when you clicked on schema, and then details, and then preview, and the web UI, that's all metadata that stored so, you're not charged for that. Also, counting like select counts star, how many records are in the table. BigQuery automatically stores the count of records, the total number of records, that's metadata, that you're not charged for as well. One of my favorite hours is those cached queries, so as you saw Wikipedia query. We ran it once, it took about 10-15 seconds to run, and then we ran it again. It's always a fun dumb one to do, returns the cash resolved. That's because you've already ran that query, and as we get into, in the course immediately after this one, and we'll sneakily reveal now every query that you run, stores a temporary table underneath in BigQuery, that's the basis of cache. Queries with errors, again there's an asterisk here. Error in the sense of a syntax error, fantastic error, as in it's human error, meaning that you ran a query, but you realize it was a wrong query, you didn't want to process so much data. If you cancel it halfway through, you could be charged potentially for up to the data that you've processed, or even the entire dataset that was quoted. Okay, so that's the three tiers of pricing. So, let's dive into a little bit more details on the pricing specifics. And again, all of this is calculated with the fact that pricing constantly changes. So, it's hard to do a lecture on pricing, but just keep these general kind of principles in mind. Queries are charged to the amount of bytes that you process, and that's largely, what we're going to focus on optimizing especially, when we get into this section of, "All right, well how do I actually reduce the amount of bytes that I'm processing?" So it's five dollars per TB, after the first TB per month because the first TB per month is free. Storage, as we mentioned is two cents per GB per month, after 90 days if your table has not had any edits to it, or additions drops down by half, and screening inserts when you're inputting in your real time, individual records to BigQuery is per cents per GB. Again the pricing link is there for you. Storage pricing is prorated. So, if it's half a month for storage again, you pay that prorated costs.
