Guia de início rápido H3 (Databricks SQL)

Artigo
12/27/2024

O início rápido das funções geoespaciais H3 nesta página ilustra o seguinte:

Como carregar conjuntos de dados de geolocalização no Unity Catalog.
Como converter latitude e longitude columns em células H3 columns.
Como converter polígono de código postal ou multipolígono WKT columns para célula H3 columns.
Como consultar a análise de embarque e desembarque do Aeroporto LaGuardia para o Distrito Financeiro de Manhattan.
Como renderizar contagens agregadas H3 em um mapa.

Exemplos de blocos de notas e consultas

Preparar dados do Unity Catalog

Neste caderno nós:

Set o conjunto de dados de táxi público do Databricks Filesystem.
Set atualize o conjunto de dados NYC Zip Code.

Preparar dados do Unity Catalog

Get caderno

Consultas SQL Databricks com Databricks Runtime 11.3 LTS e superior

Consulta 1: Verifique se os dados de base foram configurados. Consulte Bloco de notas.

use catalog geospatial_docs;
use database nyc_taxi;
show tables;
-- Verify initial data is setup (see instructions in setup notebook)
-- select format_number(count(*),0) as count from yellow_trip;
-- select * from nyc_zipcode;

Consulta 2: H3 NYC CEP - Aplicar h3_polyfillash3 na resolução 12.

use catalog geospatial_docs;
use database nyc_taxi;
-- drop table if exists nyc_zipcode_h3_12;
create table if not exists nyc_zipcode_h3_12 as (
  select
    explode(h3_polyfillash3(geom_wkt, 12)) as cell,
    zipcode,
    po_name,
    county
  from
    nyc_zipcode
);
-- optional: zorder by `cell`
optimize nyc_zipcode_h3_12 zorder by (cell);
select
  *
from
  nyc_zipcode_h3_12;

Consulta 3: H3 Taxi Trips - Aplicar h3_longlatash3 na resolução 12.

use catalog geospatial_docs;
use database nyc_taxi;
-- drop table if exists yellow_trip_h3_12;
create table if not exists yellow_trip_h3_12 as (
  select
    h3_longlatash3(pickup_longitude, pickup_latitude, 12) as pickup_cell,
    h3_longlatash3(dropoff_longitude, dropoff_latitude, 12) as dropoff_cell,
    *
  except
    (
      rate_code_id,
      store_and_fwd_flag
    )
  from
    yellow_trip
);
-- optional: zorder by `pickup_cell`
-- optimize yellow_trip_h3_12 zorder by (pickup_cell);
select
  *
from
  yellow_trip_h3_12
 where pickup_cell is not null;

Consulta 4: Captadores LGA H3 - Captadores 25M do LaGuardia (LGA)

use catalog geospatial_docs;
use database nyc_taxi;
create
or replace view lga_pickup_h3_12 as (
  select
    t.*
  except(cell),
    s.*
  from
    yellow_trip_h3_12 as s
    inner join nyc_zipcode_h3_12 as t on s.pickup_cell = t.cell
  where
    t.zipcode = '11371'
);
select
  format_number(count(*), 0) as count
from
  lga_pickup_h3_12;
-- select
  --   *
  -- from
  --   lga_pickup_h3_12;

Consulta 5: H3 Financial District Dropoffs - 34M total de dropoffs no Financial District

use catalog geospatial_docs;
use database nyc_taxi;
create
or replace view fd_dropoff_h3_12 as (
  select
    t.*
  except(cell),
    s.*
  from
    yellow_trip_h3_12 as s
    inner join nyc_zipcode_h3_12 as t on s.dropoff_cell = t.cell
  where
    t.zipcode in ('10004', '10005', '10006', '10007', '10038')
);
select
  format_number(count(*), 0) as count
from
  fd_dropoff_h3_12;
-- select * from fd_dropoff_h3_12;

Consulta 6: H3 LGA-FD - 827K dropoffs em FD com pickup de LGA

use catalog geospatial_docs;
use database nyc_taxi;
create
or replace view lga_fd_dropoff_h3_12 as (
  select
    *
  from
    fd_dropoff_h3_12
  where
    pickup_cell in (
      select
        distinct pickup_cell
      from
        lga_pickup_h3_12
    )
);
select
  format_number(count(*), 0) as count
from
  lga_fd_dropoff_h3_12;
-- select * from lga_fd_dropoff_h3_12;

Consulta 7: LGA-FD por código postal - Contagem de devoluções FD por código postal + gráfico de barras

use catalog geospatial_docs;
use database nyc_taxi;
select
  zipcode,
  count(*) as count
from
  lga_fd_dropoff_h3_12
group by
  zipcode
order by
  zipcode;

Consulta 8: LGA-FD por H3 - Contagem de devoluções FD por célula H3 + visualização do marcador de mapa

use catalog geospatial_docs;
use database nyc_taxi;
select
  zipcode,
  dropoff_cell,
  h3_centerasgeojson(dropoff_cell) :coordinates [0] as dropoff_centroid_x,
  h3_centerasgeojson(dropoff_cell) :coordinates [1] as dropoff_centroid_y,
  format_number(count(*), 0) as count_disp,
  count(*) as `count`
from
  lga_fd_dropoff_h3_12
group by
  zipcode,
  dropoff_cell
order by
  zipcode,
  `count` DESC;

LGA-FD H3 Contagens 1