The shape of a dataset is not preserved when a run is exported to another database
simonzihlmann opened this issue · 1 comments
simonzihlmann commented
As the title says, the shape of a dataset is not preserved when it is exported from one database to another one. See below for more details.
Steps to reproduce
- generate a dummy dataset
initialise_or_create_database_at("./source.db")
# Setting up a doNd measurement
sweep_1 = LinSweep(dac.ch1, -1, 1, 20, 0.01)
sweep_2 = LinSweep(dac.ch2, -1, 1, 20, 0.01)
dond(
sweep_1, # 1st independent parameter
sweep_2, # 2nd independent parameter
dmm.v1, # 1st dependent parameter
dmm.v2, # 2nd dependent parameter
measurement_name="dond_example", # Set the measurement name
exp=tutorial_exp, # Set the experiment to save data to.
show_progress=True # Optional progress bar
)
ds_source = load_by_run_spec(experiment_name='tutorial_exp', captured_run_id=1)
ds_source.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
- Extract it to another database
from qcodes.dataset.sqlite.database import connect
source_path = os.path.join(os.getcwd(), './source.db')
target_path = os.path.join(os.getcwd(), './target.db')
source_conn = connect(source_path)
target_conn = connect(target_path)
extract_runs_into_db(source_path, target_path, 1)
- load it from the new database
initialise_or_create_database_at("./target.db")
ds_target = load_by_run_spec(experiment_name='tutorial_exp', captured_run_id=1)
ds_target.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
Expected behaviour
ds_source.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
and ds_target.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
should return in both cases (20, 20)
Actual behaviour
ds_source.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
returns (20, 20)
ds_target.get_parameter_data('dmm_v1')['dmm_v1']['dmm_v1'].shape
returns (400,)
It would be nice if the shape would be preserved as it facilitates further data processing.
System
operating system
Win10
If you are using a released version of qcodes (recommended):
0.32.0
simonzihlmann commented
@jenshnielsen thanks for this rapid fix.