You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Sep 1, 2022. It is now read-only.
each file contains three-hourly time steps with the time variable being a float with units "hours since 1970-01-01 00:00:00 UTC" with the first time step in the first file (200601) equal to 315576.0 (1970-01-01 00:00:00 UTC + 315576.0 hours = 2006-01-01 00:00:00).
But the new time created by FMRC has units "hours since 2006-01-01 00:01:00.000 UTC". I was wandering where does this extra minute (00:01:00.000) comes from? Just having a slightly non stardard time units would not be a problem, but this extra minute introduces a truncation error and a query in ncss results in:
Is there a way of forcing FMRC to use a specific time:units or some other approach to correct this behavior? I am using thredds version 4.6.11.
If I join the monthly files in annual files using NCO (ncrcat) and change the collection spec to <collection spec="/home/opendap/datasets/noaa/ww3res05/**/multi_1.glo_30m.#yyyy\.nc"/>
the aggregated time units becomes "hours since 2006-01-01 00:00:00.000 UTC" and the truncation error disappear.
Thank you very much.
The text was updated successfully, but these errors were encountered:
the problem is in the expression to extract the date that you are using in the spec.
you must use upper case M for month, no lower case m which means minutes.
This error makes to interpret the 201601 date as year=2016 and minute=1, and the rest of elements of the date are interpreted as default, i.e. month and day with value 1 and time elements as 0 valued
You are completely right @cofinoa, just replacing mm for MM did the trick. My bad for trying to write by memory instead of just copying from somewhere. I lost two days trying to correct this problem (writing the time as integer, changing the units to seconds, etc) and would probably waist more time if you hadn't figured out.
Hello, I have several monthly files aggregated in a FMRC with the usual code:
<featureCollection name="noaa_ww3_prodhind" featureType="FMRC" harvest="true" path="noaa/ww3/prodhind"> <collection spec="/home/opendap/datasets/noaa/ww3res05/**/multi_1.glo_30m.#yyyymm#\.nc"/> <update startup="true"/> <fmrcConfig regularize="false" datasetTypes="Best Files" /> </featureCollection>
each file contains three-hourly time steps with the time variable being a float with units "hours since 1970-01-01 00:00:00 UTC" with the first time step in the first file (200601) equal to 315576.0 (1970-01-01 00:00:00 UTC + 315576.0 hours = 2006-01-01 00:00:00).
But the new time created by FMRC has units "hours since 2006-01-01 00:01:00.000 UTC". I was wandering where does this extra minute (00:01:00.000) comes from? Just having a slightly non stardard time units would not be a problem, but this extra minute introduces a truncation error and a query in ncss results in:
Is there a way of forcing FMRC to use a specific time:units or some other approach to correct this behavior? I am using thredds version 4.6.11.
If I join the monthly files in annual files using NCO (ncrcat) and change the collection spec to
<collection spec="/home/opendap/datasets/noaa/ww3res05/**/multi_1.glo_30m.#yyyy\.nc"/>
the aggregated time units becomes "hours since 2006-01-01 00:00:00.000 UTC" and the truncation error disappear.
Thank you very much.
The text was updated successfully, but these errors were encountered: