metadata comparison too strict for hatyan.read_dia()

[veenstrajelmer]

The code below raises "ValueError: metadata for two datasets is not equal, cannot be merged:" because of varying timestep_min. This should be ignored when merging timeseries.

import hatyan
file_dia = r"p:\archivedprojects\11208031-010-kenmerkende-waarden-k\work\data_vanRWS_20220805\wetransfer_waterstandsgegevens_2022-08-05_1306\WATHTE_10min\HOEKVHLD_1.dia"
ts = hatyan.read_dia(file_dia, block_ids="allstation", station="HOEKVHLD")
ts["values"].plot()

The code for metadata comparison is this part:

hatyan/hatyan/timeseries.py

Lines 1584 to 1590 in 8a004bc

	#concat allyears dataset
	data_pd_all = pd.concat(data_pd_list)
	metadata_compare(metadata_list)
	metadata = metadata_list[0].copy()
	metadata['tstart'] = metadata_list[0]['tstart']
	metadata['tstop'] = metadata_list[-1]['tstop']
	data_pd_all = metadata_add_to_obj(data_pd_all,metadata)

This also includes setting the metadata from the first block as the metadata of the merged timeseries, which will then not contain a timestep_min. Also the tstart/tstop are derived assuming the blocks are ordered in time which is not always the case.

Todo:

• add testcase that fails with current setup
• drop tstart/tstop attrs from timeseries dataframes, only keep for component dataframes
• drop tzone attrs from timeseries dataframes, only keep for component dataframes
• drop timestep_min attrs from timeseries dataframes, can probably be dropped throughout the code
• remove timestep_min from metadata before comparing them (but keep in mind that it is added as metadata to the merged dataset also) >> not added to metadata anymore