Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK directory-like layout is perfect for what I'm working on, for record storage, with one caveat. I'm curious if there's a way to set it up to return default records in the event a specific record isn't available (I'm attempting to reduce the number of round-trips to retrieve some data).

For example:

I create a node named "foo" with the data "lorem_ipsum":
create /foo lorem_ipsum

Then I create a node named "bar" under "foo" with the data "baz":
create /foo/bar baz

If I get "/foo/bar" obviously I'll get "baz", however, if I try to get "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly suspect as a design pattern, but there's a method to the madness. We want to create records that can be overridden by sub-records (like "/foo/bar"), but in the event that sub-record doesn't exist, I'd get back the next concrete record.

Any thoughts? Thanks!

-Ryan

ZK doesn't do this out of the box, but it is very easy to wrap up ZK
primitives to get this behavior.

On Wed, Feb 22, 2012 at 12:45 AM, Shelley, Ryan <[EMAIL PROTECTED]>wrote:

> Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK
> directory-like layout is perfect for what I'm working on, for record
> storage, with one caveat. I'm curious if there's a way to set it up to
> return default records in the event a specific record isn't available (I'm
> attempting to reduce the number of round-trips to retrieve some data).
>
> For example:
>
> I create a node named "foo" with the data "lorem_ipsum":
> create /foo lorem_ipsum
>
> Then I create a node named "bar" under "foo" with the data "baz":
> create /foo/bar baz
>
> If I get "/foo/bar" obviously I'll get "baz", however, if I try to get
> "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly suspect as
> a design pattern, but there's a method to the madness. We want to create
> records that can be overridden by sub-records (like "/foo/bar"), but in the
> event that sub-record doesn't exist, I'd get back the next concrete record.
>
> Any thoughts? Thanks!
>
> -Ryan
>

It seems like the multi-txn API might be helpful here. Since each path is
of known length, you could have a multi-op query that is

get(/a/b/c/d), get(/a/b/c), get(/a/b), get(/a)

and then check the results of each in order.

Henry

On 21 February 2012 16:45, Shelley, Ryan <[EMAIL PROTECTED]> wrote:

> Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK
> directory-like layout is perfect for what I'm working on, for record
> storage, with one caveat. I'm curious if there's a way to set it up to
> return default records in the event a specific record isn't available (I'm
> attempting to reduce the number of round-trips to retrieve some data).
>
> For example:
>
> I create a node named "foo" with the data "lorem_ipsum":
> create /foo lorem_ipsum
>
> Then I create a node named "bar" under "foo" with the data "baz":
> create /foo/bar baz
>
> If I get "/foo/bar" obviously I'll get "baz", however, if I try to get
> "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly suspect as
> a design pattern, but there's a method to the madness. We want to create
> records that can be overridden by sub-records (like "/foo/bar"), but in the
> event that sub-record doesn't exist, I'd get back the next concrete record.
>
> Any thoughts? Thanks!
>
> -Ryan
>

--
Henry Robinson
Software Engineer
Cloudera
415-994-6679

Great input folks! Thanks so much for your suggestions. I'm going to
fiddle with this a bit today and go from there. Thanks again!

-Ryan

On 2/21/12 5:19 PM, "Henry Robinson" <[EMAIL PROTECTED]> wrote:

>It seems like the multi-txn API might be helpful here. Since each path is
>of known length, you could have a multi-op query that is
>
>get(/a/b/c/d), get(/a/b/c), get(/a/b), get(/a)
>
>and then check the results of each in order.
>
>Henry
>
>On 21 February 2012 16:45, Shelley, Ryan <[EMAIL PROTECTED]> wrote:
>
>> Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK
>> directory-like layout is perfect for what I'm working on, for record
>> storage, with one caveat. I'm curious if there's a way to set it up to
>> return default records in the event a specific record isn't available
>>(I'm
>> attempting to reduce the number of round-trips to retrieve some data).
>>
>> For example:
>>
>> I create a node named "foo" with the data "lorem_ipsum":
>> create /foo lorem_ipsum
>>
>> Then I create a node named "bar" under "foo" with the data "baz":
>> create /foo/bar baz
>>
>> If I get "/foo/bar" obviously I'll get "baz", however, if I try to get
>> "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly
>>suspect as
>> a design pattern, but there's a method to the madness. We want to create
>> records that can be overridden by sub-records (like "/foo/bar"), but in
>>the
>> event that sub-record doesn't exist, I'd get back the next concrete
>>record.
>>
>> Any thoughts? Thanks!
>>
>> -Ryan
>>
>
>
>
>--
>Henry Robinson
>Software Engineer
>Cloudera
>415-994-6679

Would the multi-txn API docs be somewhere I can peruse? I haven't found
anything on the site or via Google, and the sources don't seem to have
JavaDocs on them. All I can seem to find on the usage of the Java API are
the Java examples, and those don't have an example of doing a multi-op.

-Ryan

On 2/21/12 5:19 PM, "Henry Robinson" <[EMAIL PROTECTED]> wrote:

>It seems like the multi-txn API might be helpful here. Since each path is
>of known length, you could have a multi-op query that is
>
>get(/a/b/c/d), get(/a/b/c), get(/a/b), get(/a)
>
>and then check the results of each in order.
>
>Henry
>
>On 21 February 2012 16:45, Shelley, Ryan <[EMAIL PROTECTED]> wrote:
>
>> Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK
>> directory-like layout is perfect for what I'm working on, for record
>> storage, with one caveat. I'm curious if there's a way to set it up to
>> return default records in the event a specific record isn't available
>>(I'm
>> attempting to reduce the number of round-trips to retrieve some data).
>>
>> For example:
>>
>> I create a node named "foo" with the data "lorem_ipsum":
>> create /foo lorem_ipsum
>>
>> Then I create a node named "bar" under "foo" with the data "baz":
>> create /foo/bar baz
>>
>> If I get "/foo/bar" obviously I'll get "baz", however, if I try to get
>> "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly
>>suspect as
>> a design pattern, but there's a method to the madness. We want to create
>> records that can be overridden by sub-records (like "/foo/bar"), but in
>>the
>> event that sub-record doesn't exist, I'd get back the next concrete
>>record.
>>
>> Any thoughts? Thanks!
>>
>> -Ryan
>>
>
>
>
>--
>Henry Robinson
>Software Engineer
>Cloudera
>415-994-6679

Yeah, i also find the problem of "Java Sample"
Anyone has the permissions to create some?

2012/2/23 Shelley, Ryan <[EMAIL PROTECTED]>

> Would the multi-txn API docs be somewhere I can peruse? I haven't found
> anything on the site or via Google, and the sources don't seem to have
> JavaDocs on them. All I can seem to find on the usage of the Java API are
> the Java examples, and those don't have an example of doing a multi-op.
>
> -Ryan
>
> On 2/21/12 5:19 PM, "Henry Robinson" <[EMAIL PROTECTED]> wrote:
>
> >It seems like the multi-txn API might be helpful here. Since each path is
> >of known length, you could have a multi-op query that is
> >
> >get(/a/b/c/d), get(/a/b/c), get(/a/b), get(/a)
> >
> >and then check the results of each in order.
> >
> >Henry
> >
> >On 21 February 2012 16:45, Shelley, Ryan <[EMAIL PROTECTED]> wrote:
> >
> >> Hi folks. Very new to ZooKeeper and evaluating it for a project. The ZK
> >> directory-like layout is perfect for what I'm working on, for record
> >> storage, with one caveat. I'm curious if there's a way to set it up to
> >> return default records in the event a specific record isn't available
> >>(I'm
> >> attempting to reduce the number of round-trips to retrieve some data).
> >>
> >> For example:
> >>
> >> I create a node named "foo" with the data "lorem_ipsum":
> >> create /foo lorem_ipsum
> >>
> >> Then I create a node named "bar" under "foo" with the data "baz":
> >> create /foo/bar baz
> >>
> >> If I get "/foo/bar" obviously I'll get "baz", however, if I try to get
> >> "/foo/blah" I'd like to get "lorem_ipsum." I know this is highly
> >>suspect as
> >> a design pattern, but there's a method to the madness. We want to create
> >> records that can be overridden by sub-records (like "/foo/bar"), but in
> >>the
> >> event that sub-record doesn't exist, I'd get back the next concrete
> >>record.
> >>
> >> Any thoughts? Thanks!
> >>
> >> -Ryan
> >>
> >
> >
> >
> >--
> >Henry Robinson
> >Software Engineer
> >Cloudera
> >415-994-6679
>
>