html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue
https://github.com/pydata/xarray/issues/597#issuecomment-370381376,https://api.github.com/repos/pydata/xarray/issues/597,370381376,MDEyOklzc3VlQ29tbWVudDM3MDM4MTM3Ng==,3404817,2018-03-05T10:49:19Z,2018-03-05T10:49:19Z,CONTRIBUTOR,Can't this be closed?,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,109202603
https://github.com/pydata/xarray/issues/597#issuecomment-144794506,https://api.github.com/repos/pydata/xarray/issues/597,144794506,MDEyOklzc3VlQ29tbWVudDE0NDc5NDUwNg==,1217238,2015-10-01T17:30:55Z,2015-10-01T17:30:55Z,MEMBER,"So unfortunately there isn't an easy way to handle irregular data like this with xray. Before you put this stuff in a single dataset, you would need to align the time variables, probably by doing `.reindex` with the [`method` argument](http://xray.readthedocs.org/en/stable/indexing.html#nearest-neighbor-lookups), which lets you do nearest-neighbor interpolation, or by using [`.resample`](http://xray.readthedocs.org/en/stable/time-series.html#resampling-and-grouped-operations). That way, you could make, for example, a dataset with the latest image for each location at the start of each month.

The other option (probably also useful) is to only concatenate one spatial tile along time, so you don't need to do any interpolation or resampling. The should work out of the box with `open_mfdataset`.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,109202603
https://github.com/pydata/xarray/issues/597#issuecomment-144680733,https://api.github.com/repos/pydata/xarray/issues/597,144680733,MDEyOklzc3VlQ29tbWVudDE0NDY4MDczMw==,2526498,2015-10-01T09:49:11Z,2015-10-01T09:59:56Z,NONE,"I have created the NetCDF files myself from geotiffs and I have made them so that there is no geographical overlapping between them. Basically each file contains a 1x1 degree area and a year worth of satellite data. The only problem that I can see with this approach is that the time dimension is different between files (the satellite covers different areas at different times). This might be problem when aggregating a big area because if the time dimension has to be homogenised it will be filled with basically no data over the whole area (sparse arrays). Depending how this sparsity is implemented it can fill memory pretty quickly. 

Some of these files can be found at:
http://dapds00.nci.org.au/thredds/catalog/uc0/rs0_dev/gdf_trial/20150709/LS5TM/catalog.html

A sample of `open_dataset` for a file is:

```
import xray
dap_file = 'http://dapds00.nci.org.au/thredds/dodsC/uc0/rs0_dev/gdf_trial/20150709/LS5TM/LS5TM_1987_-34_147.nc'
ds = xray.open_dataset(dap_file, decode_coords=False)
print(ds)

<xray.Dataset>
Dimensions:    (latitude: 4000, longitude: 4000, time: 11)
Coordinates:
  * time       (time) datetime64[ns] 1987-05-27T23:26:36 1987-08-31T23:29:21 ...
  * latitude   (latitude) float64 -33.0 -33.0 -33.0 -33.0 -33.0 -33.0 -33.0 ...
  * longitude  (longitude) float64 147.0 147.0 147.0 147.0 147.0 147.0 147.0 ...
Data variables:
    crs        int32 ...
    B10        (time, latitude, longitude) float64 ...
    B20        (time, latitude, longitude) float64 ...
    B30        (time, latitude, longitude) float64 ...
    B40        (time, latitude, longitude) float64 ...
    B50        (time, latitude, longitude) float64 ...
    B70        (time, latitude, longitude) float64 ...
Attributes:
    history: NetCDF-CF file created 20150709.
    license: Generalised Data Framework NetCDF-CF Test File
    spatial_coverage: 1.000000 degrees grid
    featureType: grid
    geospatial_lat_min: -34.0
    geospatial_lat_max: -33.0
    geospatial_lat_units: degrees_north
    geospatial_lat_resolution: -0.00025
    geospatial_lon_min: 147.0
    geospatial_lon_max: 148.0
    geospatial_lon_units: degrees_east
    geospatial_lon_resolution: 0.00025
```

Thank you very much for your help.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,109202603
https://github.com/pydata/xarray/issues/597#issuecomment-144577690,https://api.github.com/repos/pydata/xarray/issues/597,144577690,MDEyOklzc3VlQ29tbWVudDE0NDU3NzY5MA==,1217238,2015-09-30T23:59:04Z,2015-09-30T23:59:04Z,MEMBER,"It would probably be helpful to show the results of printing the several of these datasets when opened via `open_dataset`.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,109202603
https://github.com/pydata/xarray/issues/597#issuecomment-144577524,https://api.github.com/repos/pydata/xarray/issues/597,144577524,MDEyOklzc3VlQ29tbWVudDE0NDU3NzUyNA==,1217238,2015-09-30T23:57:38Z,2015-09-30T23:57:38Z,MEMBER,"What do these different time range and geographical regions look like?

If they are adjacent and non-overlapping, then xray could be a very good fit and I can help whip up an example to get you started. If this is not the case (and I know that can be an issue with satellite data), then it's going to be more awkward to put them together into a single logical Dataset. It might make more sense to work with individual Datasets, e.g., at the level of an individual satellite image.
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,109202603