Why it‘s so slow when concat 2 dataarrays？

[forestbat]

I have two DataArrays like below：

<xarray.DataArray (variable: 5, basin_id: 601, time: 23121)> Size: 556MB
Coordinates:
  * basin_id  (basin_id) <U25 60kB 'camels_12117000' ... 'WI_DCP_SOSW3'
  * time      (time) datetime64[ns] 185kB 2016-01-01T03:00:00 ... 2023-11-30T...
  * variable  (variable) object 40B 'total_precipitation_hourly' ... 'sm_surf...

<xarray.DataArray (variable: 30, basin_id: 601, time: 23121)> Size: 3GB
Coordinates:
  * basin_id  (basin_id) <U17 41kB 'AL_DCP_HAKA1' ... 'songliao_21401900'
  * time      (time) datetime64[ns] 185kB 2016-01-01T01:00:00 ... 2023-11-30T...
  * variable  (variable) object 240B 'streamflow_up_0' ... 'streamflow_up_29'

Now I use

import xarray as xr
test_da = xr.concat([data_array1, data_array2], dim=['basin_id', 'time', 'variable'])

However this process will cost long time, so I want to know why it will become so slow and what's the correct method?

[shoyer]

Try writing xr.concat([data_array1, data_array2], dim='variable'). The syntax you used before (dim=['basin_id', 'time', 'variable']`) is not valid.