Removing location textual fields indices from tables of db_views.py file

[atalyaalon]

That's for the next milestone, but I want to keep documentation of this issue as we spoke about it in our last meeting

We can first:

• Remove indices and make sure it does not harm the cache performance
  This includes removing text fields of yishuv_name and street1_hebrew and street2_hebrew and road_segment_name from the indices of the 5 DB tables markers_hebrew, involved_markers_hebrew, vehicles_hebrew, involved_hebrew if exist in them

And then:

• discuss creating smaller tables (idea that was discussed in the past and was not implemented yet)

[ziv17]

Hi @atalyaalon
We want to add the numeric indexes if they do not exist, right? (e.g. involved_markers_hebrew does not have street2 index.

[atalyaalon]

@ziv17 yes you are right.
And in table involved_markers_hebrew we also want to remove the street1_hebrew, street1_hebrew, accident_yishuv_name , road_segment_name, as you performed in this pr

[ziv17]

Hi @atalyaalon ,
The 5 tables that are created in db_views.py are:

• AccidentMarker, or markers
• VehiclesView, or vehicles_hebrew
• Involved, or involved
• InvolvedView, or involved_hebrew
• Vehicle, or vehicles

Are these the tables that I should work on?

[ziv17]

Hi @atalyaalon ,
I might have asked you this already. If so, I apologize 😄
Currently the infographics API can accept text field as filters, and it replaces them with the corresponding numeric field. For this, it does a select with the given text field. If we keep this functionality, do we still want to remove the text indexes?

[atalyaalon]

@ziv17 can you add the like to the relevant code that does a select with the given text field?

[atalyaalon]

@ziv17 just a reminder, can you add the references to the locations we use the text fields as filters? I want to remove the text indices and replace them with numeric ones, but want to make sure this won't cause any performance issues in advance

[atalyaalon]

@ziv17 FYI I opened this issue so FE will only query by numeric fields in urban API (for road_segment API the FE already uses the API with road_segment_id param