Modify data included in the loaded gmf shapefile

[CatalinaYepes]

Currently, for loading a layer of Ground Motion Field, users can select across:

• Realization (m number of realizations)
• IMT (n number of IMTs)
• Event_ID (x number of events)
  A tmp shapefile is loaded with the following characteristics:
  Name: scenario_gmfs_rlz-00m_eid-x (for realization m and event_id x)
  Attribute table: [ IMT_1-x, IMT_2-x, ..., IMT_n-x] (for n IMTs and the event_id x)

With this approach is not possible to compare in the same layer the difference in the gmf for different event_id. Moreover, to download all the data, the user will have to upload as many layers as the number of events, which is generally a large number (for example, 1000), times the number of realizations.

To simplify the outputs, the recommendation for the tmp shapefile will be:
Name: gmfs_rlz-name_IMT-n (rlz-name is the name of the GMPE #410 )
Attribute table: [ eid-1, eid-2, ..., eid-x] (for all the events)

The new format will also allow to easier style the layer, since all the columns will have the same name across different GMPEs and IMTs.

[ptormene]

The main drawback I see in this approach is that, if the number of events can be huge, the resulting layer would have a huge number of columns.

[ptormene]

For instance, I can create a layer with 10 thousand columns (and it works), but when I attempt to open its attribute table, QGIS remains frozen for dozens of seconds on my laptop (even with 0 rows). So, I would say 10000 is feasible, but I would consider it already a upper limit. If we can have a million of events, this is completely unfeasible. The Risk and Hazard scientific teams should discuss about this and decide how to proceed.

[ptormene]

After closing the huge attribute table with 10000 fields, my QGIS froze again for some reason (for several minutes!). Therefore, it looks like even 10000 is too much.

[CatalinaYepes]

We need to think again on this issue, specially for Scenario calculations.

Can we separate the visualization? Other option will be to limit the columns to a given number of events (for example 2000?) and open a warning windows indicating that the data has been truncated.