[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fwd: Pathfinder AVHRR aggregation -- monotonic time axis



Hi Dan:

Can you try changing the coordinate axis type to double and see what happens:

       <variable name="time" type="double" >
          <attribute name="_CoordinateAxisType" value="Time" />
          <attribute name="units" value="hours since 1970-01-01 00:00:00" />
       </variable>

thanks

Daniel Holloway wrote:
John,

I'm not sure if the problem is in the TDS but thought you'd like to know. All the clients I've tried seem to indicate that the float32 values are of less precision than the originating XML indicates but I'm still double-checking our code. I'm attaching the catalog.xml for this particular dataset to this message in case you'd like to see that. The problem is potentially that the TDS is not maintaining the precision of 'Join-new' coordinate values.

------------------------------------------------------------------------



Begin forwarded message:

*From: * Daniel Holloway <address@hidden <mailto:address@hidden>>
*Date: * July 14, 2006 11:13:02 AM EDT
*To: * Steve Hankin <address@hidden <mailto:address@hidden>>
*Cc: * Ansley Manke <address@hidden <mailto:address@hidden>>, Jing Y Li <address@hidden <mailto:address@hidden>>
*Subject: * *Re: Pathfinder AVHRR aggregation -- monotonic time axis*



On Jul 13, 2006, at 7:11 PM, Steve Hankin wrote:

Hi Dan,

It turns out that the time axis has a repeated value in it Look at the coordinates with the OPeNDAP HTML (browser) interface I get

4760:1:4770

    Dataset { Float32 time[time = 11]; }
    Pathfinder/NorthwestAtlantic/nwAtlantic1KmRaw.nc;
    --------------------------------------------- time[11] 173694.38,
    173696.11, 173705.8, *173707.5, 173707.5*, 173718.22, 173719.92,
    173729.62, 173731.33, 173742.05, 173743.75

Is this a valid repeating of time points that we should learn to deal with in our software? Or a probelm in the aggregation? (There may be other repeated values. Ferret only reports the first one.)


Steve,

I've looked at the actual THREDDS xml describing the aggregation, here are the values of the 'time' coordinate around that point:
-------------


   173705.801944444    http:..../i89198174807.hdf
   173707.495555556    http:..../i89298192944.hdf
   173707.503333333    http:..../i89298193012.hdf
   173718.215  http:.../i89299061254.hdf

------------

So, the aggregations have a greater resolution than you're showing. The problem is either in the server truncating these values or on the client side. I'll double-check the server and get back to you.

   Dan


- steve

(sorry for the trouble ...)

--

º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤º¤ø,¸¸,ø¤º

Steve Hankin, NOAA/PMEL -- address@hidden <mailto:address@hidden> 7600 Sand Point Way NE, Seattle, WA 98115-6349
ph. (206) 526-6080, FAX (206) 526-6744


**
*From: * Jing Yang Li <address@hidden <mailto:address@hidden>>
*Date: * July 13, 2006 5:02:24 PM EDT
*To: * Steve Hankin <address@hidden <mailto:address@hidden>>
*Cc: * Daniel Holloway <address@hidden <mailto:address@hidden>>, Dan Holloway <address@hidden <mailto:address@hidden>>, Jonathan Callahan <address@hidden <mailto:address@hidden>>, address@hidden <mailto:address@hidden>
*Subject: * *Re: Pathfinder AVHRR aggregation -- monotonic time axis*



The correct URL is http://dods.gso.uri.edu:8080/thredds/Pathfinder/NorthwestAtlantic.html.
Jing


Steve Hankin wrote:

Hi Dan,

At a new URL now, maybe?

At http://dods.gso.uri.edu:8080/thredds I click on "Northwest Atlantic/ <Pathfinder/NorthwestAtlantic.html>" and get


HTTP Status 404 - /Pathfinder/NorthwestAtlantic.html

------------------------------------------------------------------------

    *type* Status report

    *message* _/Pathfinder/NorthwestAtlantic.html_

    *description* _The requested resource
    (/Pathfinder/NorthwestAtlantic.html) is not available._

------------------------------------------------------------------------


Apache Tomcat/5.0.28

Theories?

    - steve

=========================================================

Daniel Holloway wrote:

Hi Guys,

   I have the server running the aggregation back online.

   Dan

On Jun 27, 2006, at 7:55 PM, Steve Hankin wrote:

Hi Dan,

The AVHRR aggregation server isn't responding.  Any insights?

   - steve

=======================================================

Dan Holloway wrote:

Hi Guys,

I've installed an updated TDS revision at http:// dods.gso.uri.edu:8080/thredds that should resolve several problems we had. It resolves the problem using strides in an array subset operation, and it now allows me to add attributes. I've added several to 'lat', 'lon', and 'time', if these attributes are incorrect or you'd like something different please let me know.

Also, please let me know if there are any remaining problems with the 'time' dimension. I fixed a problem with it that showed up as a non-monotonic axis dimension but there could be other problems remaining.

   Thanks,

         Dan

p.s. I only have the Northwest Atlantic aggregations configured, I can work on the Northwest/east Pacific aggregations as well.


Jing Yang Li wrote:

Dan Holloway wrote:

Hi Jonathan, Jing,

I recently installed an updated version of the TDS, that was supposed to resolve the stride bug unfortunately it's failing at the moment. This is just a heads up that the aggregation server at URI might return errors or be unavailable over the next few days as we work thru these problems. We have to get the stride problem fixed, as well as the update that will allow me to add COARDS-like attributes to the variables in the aggregation. I will send another message once the server has been fixed and is available for you to work with. I mentioned in a previous message that I believe I resolved the non-monotonic axis issue.

    Thanks,

            Dan

Jonathan Callahan wrote:

Dan,

It turns out there are more problems with the time axis in the AVHRR aggregation. The time axis you are generating is not monotonic -- there are several places where it jumps around.

I've attached an image that plots a portion of the time aixs. The bottom axis is the time index.
REASON time axis prlblem
Ferret, along with most other packages, needs a monotonic time axis.


I can identify specific indexes or times where the jumps are occurring if that would be helpful but there may be issues with repeat values, etc. Let me know what information I can provide to help you figure out this problem.


-- Jon

Hi Dan,
I tried your new data. Yes, the monotonic problem for time axis was gone.
The time axis also passed our Java code without throwing out exceptions,
although there is still a warning like:
*** NOTE: Coordinates out of order or missing on axis time at subscript 4766.
Thanks.
Jing





-- --

Steve Hankin, NOAA/PMEL -- address@hidden <mailto:address@hidden>
7600 Sand Point Way NE, Seattle, WA 98115-0070
ph. (206) 526-6080, FAX (206) 526-6744




-- º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤º¤ø,¸¸,ø¤º

Steve Hankin, NOAA/PMEL -- address@hidden <mailto:address@hidden> 7600 Sand Point Way NE, Seattle, WA 98115-6349
ph. (206) 526-6080, FAX (206) 526-6744









NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.