Freigeben über


H3-Schnellstart (Databricks SQL)

Der Schnellstart zu räumlichen H3-Funktionen auf dieser Seite veranschaulicht Folgendes:

  • Laden des Geolocation-Datasets in den Unity Catalog.
  • Konvertieren von Breiten- und Längengradspalten in H3-Zellenspalten.
  • Konvertieren von Postleitzahl-Polygon- oder -Multipolygon-WKT-Spalten in H3-Zellenspalten.
  • Abfragen einer Abhol- und Rücktransportanalyse vom Flughafen LaGuardia zum Finanzdistrikt von Manhattan.
  • Rendern von H3-Aggregatzahlen auf einer Karte.

Beispielnotebooks und -abfragen

Vorbereiten von Unity Catalog-Daten

In diesem Notebook:

  • Einrichten des Datasets für öffentliche Taxis über Databricks Filesystem.
  • Einrichten des Datasets NYC Zip Code.

Vorbereiten von Unity Catalog-Daten

Notebook abrufen

Databricks SQL-Abfragen mit Databricks Runtime 11.3 LTS und höher

Abfrage 1: Überprüfen, ob Basisdaten eingerichtet wurden. Weitere Informationen finden Sie unter Notebook.

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;

Abfrage 2: H3-NYC-Postleitzahlcode: Anwenden von h3_polyfillash3 bei der Auflösung von 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;

Abfrage 3: H3-Taxifahrten: Anwenden von h3_longlalash3 bei der Auflösung von 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;

Abfrage 4: H3-LGA-Abholungen: 25 Millionen Abholungen von 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;

Abfrage 5: H3-Fahrten mit Ziel Finanzdistrikt: 34 Millionen Fahrten mit Ziel Finanzdistrikt

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;

Abfrage 6: H3 LGA-FD: 827.000 Fahrten mit Ziel Finanzdistrikt mit Abholung von 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;

Abfrage 7: LGA-FD nach Postleitzahl: Anzahl Fahrten mit Ziel Finanzdistrikt nach Postleitzahl und Balkendiagramm

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;

Abfrage 8: LGA-FD nach H3: Anzahl Fahrten mit Ziel Finanzdistrikt nach Zelle H3 und Visualisierung vom Typ Kartenmarkierung

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-Anzahl 1

LGA-FD H3-Anzahl 2

Notebooks für Databricks Runtime 11.3 LTS und höher

Quickstart-Python: H3-NYC-Taxi von LaGuardia nach Manhattan

Notebook abrufen

Gleiche Schnellstartstruktur wie in Databricks SQL unter Verwendung von Spark Python-Bindungen in Notebooks und kepler.gl.

Quickstart-Scala: H3-NYC-Taxi von LaGuardia nach Manhattan

Notebook abrufen

Gleiche Schnellstartstruktur wie in Databricks SQL unter Verwendung von Spark Scala-Bindungen in Notebooks und kepler.gl über Python-Zellen.

Quickstart-SQL: H3-NYC-Taxi von LaGuardia nach Manhattan

Notebook abrufen

Gleiche Schnellstartstruktur wie in Databricks SQL unter Verwendung von Spark QL-Bindungen in Notebooks und kepler.gl über Python-Zellen.