petascope time range seems to be corrupted

[joachim.ungar@…]

certain queries affecting a selection along the time axis of a XYT cube throw the following error:

Error: Could not evaluate rasdaman query: 'select csv((float)(avg_cells(((((((© [*:*,*:*,150190])>(99)))*(© [*:*,*:*,150190])))-(((((© [*:*,*:*,150190])>(200)))*(© [*:*,*:*,150190]))))))) from FSC_Scandinavia AS c'

rasdaman.log indicates that a different time range may be defined in the petascope metadatabase:

Request: 'select csv((float)(avg_cells(((((((© [*:*,*:*,150190])>(99)))*(© [*:*,*:*,150190])))-(((((© [*:*,*:*,150190])>(200)))*(© [*:*,*:*,150190]))))))) from FSC_Scandinavia AS c'…parsing…checking semantics…evaluating…r_Sinterval::calc_intersection(150190:150190, 15019000:15019048) not an interval
Error: QtVariable::evaluate() - Specified domain does not intersect with spatial domain of MDD.

This looks like an error I had after inserting a time series of images via rasgeo. Using the shift parameter "-shift 0:0:150190" in rasgeo should set the z/t axis value for the first image to 150190, for the second image to 150191 and so on. In fact it inserted in ps_domain "150190.00" for numlo and "150190.48" in numhi.

Piero helped me to modify the respective entry in ps_domain:

petascopedb=# select * from ps_domain;

id | coverage | i | name | type | numlo | numhi | strlo | strhi | uom

---

2 | 1 | 1 | y | 2 | 56.84922847 | 71.57922847 | | |
1 | 1 | 0 | x | 1 | -0.93136303 | 33.27863697 | | |
3 | 1 | 2 | t | 6 | 150190 | 150238 | | |

So, obviously, at least a part of the metadata should now be ok. Which part am I missing here?

Thanks, Joachim

[Piero Campalani]

In rasgeo you should set a cell resolution in the time domain to exactly 1, when importing a set of images.
From the README file of rasgeo:

    -csz:   z-axis cell size; if omitted, rasimport assumes x-, y-,
            and z-cell sizes are identical!

Does this fix the problem?

[joachim.ungar@…]

Great, that solved the problem. Cheers, Joachim