Remove Time Range Reading Bottleneck

[TonioF]

When a chunk store is created, the time ranges are determined. This is fast when there is a regular temporal frequency (day, month, year, ...), but in other cases (satellite-orbit-frequency, unknown), it is necessary to send multiple opensearch requests or even open the opendap data. This often slows down the opening process to more than a minute, where most time is spent on determining the time ranges.

[TonioF]

Put on hold. After a bit more investigation, it seems that this problem cannot be solved without changes to the odp.