Update session_events.py (#69)

timestamp in the data source is stored in unix epoch seconds.
tecton-ai · Dec 13, 2023 · bbc7636 · bbc7636
1 parent 4fb05ab
commit bbc7636
Showing 1 changed file with 2 additions and 1 deletion.
diff --git a/recsys/data_sources/session_events.py b/recsys/data_sources/session_events.py
@@ -6,11 +6,12 @@ def sessions_data_source_function(spark):
     schema = StructType([
         StructField("session", IntegerType(), True),
         StructField("aid", IntegerType(), True),
-        StructField("ts", TimestampType(), True),
+        StructField("ts", IntegerType(), True),
         StructField("type", StringType(), True)
     ])
 
     df = spark.read.schema(schema).parquet("s3://tecton.ai.public/tutorials/recsys-demo/train.parquet")
+    df = df.withColumn("ts", df.ts.cast(TimestampType()))
     return df
 
 sessions_batch = BatchSource(