From a6cbe3bc292362531d5bcd9305cdd019f415e988 Mon Sep 17 00:00:00 2001 From: mmaurostoffel <166130318+mmaurostoffel@users.noreply.github.com> Date: Thu, 9 Jan 2025 15:07:38 +0100 Subject: [PATCH] update des etl_capacities_weekdays.py --- etl/src/data/etl_property_capacities_weekdays.py | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/etl/src/data/etl_property_capacities_weekdays.py b/etl/src/data/etl_property_capacities_weekdays.py index 96e821d..56972ab 100644 --- a/etl/src/data/etl_property_capacities_weekdays.py +++ b/etl/src/data/etl_property_capacities_weekdays.py @@ -8,6 +8,7 @@ d = data.load() def property_capacities_weekdays(id: int, scrapeDate: str): extractions = d.extractions_propId_scrapeDate(id, scrapeDate).pl() + weekdays = ['Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday', 'Sunday'] df_calendar = pl.DataFrame() numWeeks = 0 @@ -19,12 +20,15 @@ def property_capacities_weekdays(id: int, scrapeDate: str): df_calendar = df_calendar.with_columns(pl.Series(name="dates", values=columnTitles)) df_calendar = df_calendar.with_columns((pl.col("dates").str.to_date())) numWeeks = round((df_calendar.get_column("dates").max() - df_calendar.get_column("dates").min()).days / 7, 0) - df_calendar = df_calendar.with_columns(pl.col("dates").dt.weekday().alias("weekday")) + df_calendar = df_calendar.with_columns(pl.col("dates").dt.weekday().alias("weekday_num")) + df_calendar = df_calendar.with_columns(pl.col("dates").dt.strftime("%A").alias("weekday")) df_calendar = df_calendar.drop("dates") - df_calendar = df_calendar.group_by("weekday").agg(pl.col("column_0").sum()) - df_calendar = df_calendar.with_columns((pl.col("column_0") / numWeeks).alias("weekday")) + df_calendar = df_calendar.group_by(["weekday", "weekday_num"]).agg(pl.col("column_0").sum()) + df_calendar = df_calendar.with_columns((pl.col("column_0") / numWeeks).alias("column_0")) + df_calendar = df_calendar.sort('weekday_num') + df_calendar = df_calendar.drop('weekday_num') + print(df_calendar) + result = {"scraping-date": scrapeDate, "weekday": df_calendar['weekday'].to_list(), 'capacities': df_calendar['column_0'].to_list()} - return result - - + return result \ No newline at end of file