Is this just for select queries ?

[Mahesh Velaga]

Hi,

If I understood it correctly, this is just for select queries right ?
(no transactional querying). Will there be additional operations like
update,delete etc be available in future (if not already present) ?

[Amit Agarwal]

Hi Mahesh,

BigQuery is just for select queries. We currently dont have any plans for additional operations like update, delete etc. at a transaction level.

thanks

Amit

[Michael Sheldon]

Who is mahesh...@gmail.com?

I guess we need to be careful about how we answer questions... So far at Google I've always just answered assuming everyone is a Googler. Now we'll have to check to TO line and FROM line to see how we need to answer a question.

Good learning for all!

--Michael Sheldon

[Alex Moore]

In our case we have a data set which increases at the rate of about 52
million records per year, and 150,000 per day (updated twice per
day). If I uploaded the initial data set is it possible to append
additional data? Update and delete would be awesome though (I can
think of other uses).

Additionally is it possible to pivots on data using this query
language? I don't have an account (sad face) and I couldn't work it
out from the docs.

On May 20, 6:22 pm, Michael Sheldon <mshel...@google.com> wrote:
> Who is maheshraj...@gmail.com?

>
> I guess we need to be careful about how we answer questions... So far at
> Google I've always just answered assuming everyone is a Googler. Now we'll
> have to check to TO line and FROM line to see how we need to answer a
> question.
>
> Good learning for all!
>
> --Michael Sheldon
>

> On Thu, May 20, 2010 at 1:26 AM, Amit Agarwal <amit.agar...@gmail.com>wrote:
>
>
>
> > Hi Mahesh,
>
> > BigQuery is just for select queries. We currently dont have any plans for
> > additional operations like update, delete etc. at a transaction level.
>
> > thanks
> > Amit
>

[Michael Sheldon]

Hello Alex,

Thank you for the interest in BigQuery. Let me try and answer your questions:

Yes, you can append additional records multiple times a day to your BigQuery Tables. (We intend to have a "continuous import" feature to make this easier by letting you configure your BigQuery Table to automatically import all data you push into your Google Storage for Developers bucket. However that feature is a bit further out.)

Our query language does not support Pivots. Thanks for asking: we'll take that as a feature request.

Regarding accounts:

• You can start by making sure you have a Google Storage for Developers account. You can find the sign up link for this service at http://code.google.com/apis/storage/.
• You can find the BigQuery sign up link near the bottom of http://code.google.com/apis/bigquery/. We had more interest in our service than we anticipated, so it will take a while before we can provision accounts for everyone interested. We will get back to everyone that signs up. Thanks for your patience!

Cheers,

--Michael Sheldon

On Sat, May 22, 2010 at 4:45 PM, Alex Moore <zyc...@gmail.com> wrote:

In our case we have a data set which increases at the rate of about 52
million records per year, and 150,000 per day (updated twice per
day).   If I uploaded the initial data set is it possible to append
additional data?  Update and delete would be awesome though (I can
think of other uses).

Additionally is it possible to pivots on data using this query
language?  I don't have an account (sad face) and I couldn't work it
out from the docs.

[kazunori_279]

Hi Micheal,

It was so amazing to see that the queries for 60B records are done in
a few seconds.
What kind of approach do you use inside it?

a. on-demand massively parallel query with thousands of nodes
b. pre-aggregation/computation (like building B-Trees for all the
columns) at the import time
c. any other unknown technique

And how long did it take to import the 60B records?

Thanks,

Kaz

On 5月26日, 午前3:52, Michael Sheldon <mshel...@google.com> wrote:
> Hello Alex,
>
> Thank you for the interest in BigQuery. Let me try and answer your
> questions:
>
> Yes, you can append additional records multiple times a day to your BigQuery
> Tables. (We intend to have a "continuous import" feature to make this easier
> by letting you configure your BigQuery Table to automatically import all
> data you push into your Google Storage for Developers bucket. However that
> feature is a bit further out.)
>
> Our query language does not support Pivots. Thanks for asking: we'll take
> that as a feature request.
>
> Regarding accounts:
>

>    - You can start by making sure you have a Google Storage for Developers

>    account. You can find the sign up link for this service at
>    http://code.google.com/apis/storage/.

>    - You can find the BigQuery sign up link near the bottom of

>    http://code.google.com/apis/bigquery/. We had more interest in our
>    service than we anticipated, so it will take a while before we can provision
>    accounts for everyone interested. We will get back to everyone that signs
>    up. Thanks for your patience!
>
> Cheers,
>
> --Michael Sheldon
>

[Alex Moore]

Yes, 60B is what really drew my attention to it :)

Will BigQuery be available for "almost" real time reporting?
How long will it take (i.e. is it planned to take) for the data to be
"query-able" in BigQuery after the data has been uploaded to google
storage?

Any ideas on what the pricing model is going to be like? Per query,
per amount of data?

[Amit Agarwal]

Kaz,

Thank you for the interest in BigQuery.

We don't comment on the backend details  of BigQuery. What I can say is that we are running BigQuery on thousands of machines.

The 60 B records were imported as part of a sample dataset. We don't have numbers to share at this point regarding the time taken to import the dataset. 

thanks

Amit

[Amit Agarwal]

Thanks Alex. Comments inline.

Our first priority is get the ability to import data into BigQuery. 

On Wed, May 26, 2010 at 4:42 PM, Alex Moore <zyc...@gmail.com> wrote:

Yes, 60B is what really drew my attention to it :)

Will BigQuery be available for "almost" real time reporting?
How long will it take (i.e. is it planned to take) for the data to be
"query-able" in BigQuery after the data has been uploaded to google
storage?

Real time has different meanings for different users and different use cases. We are working on making the data "query-able" as fast as possible once it has been uploaded to Google Storage for developers.

Any ideas on what the pricing model is going to be like? Per query,
per amount of data?

Currently BigQuery is in preview and we don't have anything to share on pricing.  During this preview period we want to understand and tune the system before we work on pricing.

thanks

Amit