| cd /home/cloudera/Downloads | |
| hadoop fs -ls / | |
| hadoop fs -mkdir -p /data/baseball/team/ | |
| hadoop fs -copyFromLocal baseball/team.csv /data/baseball/team/ | |
| CREATE EXTERNAL TABLE team ( | |
| year string, | |
| league_id string, | |
| team_id string, |
| DataFrame zoneDF = spark.Read() | |
| .Option("header","true") | |
| .Schema(taxiZoneSchema) | |
| .Csv(taxiZoneSourcePath); | |
| zoneDF.Write().Format("delta").Mode("overwrite").Save(taxiZonePath); | |
| zoneDF.Show(); |
| val zoneDF = spark.read.option("header","true").schema(taxiZoneSchema).csv(taxiZoneSourcePath) | |
| zoneDF.write.format("delta").mode("overwrite").save(taxiZonePath) | |
| zoneDF.show() |
| from pyspark.sql.functions import col, desc, regexp_replace, substring, to_date, from_json, explode, expr | |
| from pyspark.sql.types import StructType, StringType | |
| yellow_source_path = "wasbs://nyctlc@azureopendatastorage.blob.core.windows.net/yellow/puYear=2018/puMonth=*/*.parquet" | |
| taxi_zone_source_path = "abfss://demo@datakickstartadls.dfs.core.windows.net/nyctaxi/lookups/taxi_zone_lookup.csv" | |
| taxi_zone_path = "abfss://demo@datakickstartadls.dfs.core.windows.net/nyctaxi/lookups/taxi_zone" | |
| taxi_rate_path = "abfss://demo@datakickstartadls.dfs.core.windows.net/nyctaxi/lookups/taxi_rate_code" | |
| yellow_delta_path = "abfss://demo@datakickstartadls.dfs.core.windows.net/nyctaxi/tripdata/yellow_delta" |
| // Implicits provide many shortcuts, including conversion from Row into a specific type | |
| import spark.implicits._ | |
| // Case class to use as type for each Row | |
| case class VehicleStopRaw( | |
| stop_id: String, stop_cause: String, service_area: String, subject_race: String, | |
| subject_sex: String, subject_age: String, timestamp: String, stop_date: String, | |
| stop_time: String, sd_resident: String, arrested: String, searched: String, | |
| obtained_consent: String, contraband_found: String, property_seized: String) |
https://docs.ksqldb.io/en/latest/concepts/tables/ "A table is a mutable, partitioned collection that models change over time...represents what is true as of 'now'"
A table can have a primary key that handles updates and deletes based on that primary key which makes them very useful for accurate aggregations when data may change over time.