The template dataset should be a subset of the dataset being flagged, and the order of the records in the two datasets should be the same. Normally correlations in the template and the input are matched by comparing time, polarisation, baseline and frequency. However options nopol and nofreq can turn off the matching of polarisation and frequency. Typical inputs might be:
|Dataset to be flagged.
|Template dataset, say a channel-0 one.
|Do not match on frequency,
|i.e. ignore the frequency in multi.chan0