I am using standard SQL and am trying to add the weekly sum for product usage by week.
Using code below, I was able to add to each row the respective week and year it falls into. How would I go about summing the totals for an item by week and outputting it in columns, say up to the last 8 weeks.
extract(week from Metrics_Date) as week, EXTRACT(YEAR FROM Metrics_Date) AS year
Image is my raw data with the week and year next to an item:
This image is of above raw data being analyzed further(grouping them together). Here is where I would want to add columns, current_week & firstday of week date, and a sum of that weeks totals.
Any help would be appreciated.
You don't need the extract() by the way, you can do truncation DATE_TRUNC(your_date, WEEK) and it will truncate it to the week, usually easier.
Also, because the result of the truncation is a date, you will have the first day of the week already.
The rest I believe you have it figured out already, but just in case:
SELECT DATE_TRUNC(your_date_field, WEEK) AS week, SUM(message_count) AS total_messages FROM your_table GROUP BY 1
Related
I have a dataset with the structure below.
I want to calculate a monthly average of the views.
I attempted to calculate the yearly frequency with the following code and I believe it is correct
SELECT
EXTRACT (YEAR FROM TO_DATE("date",'Month YYYY') ) AS "year",
AVG("views")
FROM talks
GROUP BY EXTRACT (YEAR FROM TO_DATE("date",'Month YYYY') )
ORDER BY "year" DESC
When it comes to the monthly analysis I have the problem that there several records for the same month in a year and there several years with the same months (in reality the dataset has information for many years - this a reduced version).
How can I go to implement this?
If you want the average per month then just group by your current date field.
If you want the average per month regardless of year then you would have to extract the month part of the current date field and group by that.
But your date field now appears to be having string data type; it would be better to use proper date data type. Then your analysis would be much easier, more flexible, better performing.
Scenario: From bigquery, have to fetch the specified date's week data + its previous week data + its next future week data. Week starts is Wednesday.
Tried Query:
Select * from table
and extract(week(wednesday) from Calendar_Day) >= (extract(week(wednesday) from PARSE_DATE('%d/%m/%Y','21/10/2020')) - 1)
and extract(week(wednesday) from Calendar_Day) >= (extract(week(wednesday) from PARSE_DATE('%d/%m/%Y','21/10/2020') ))
and extract(week(wednesday) from Calendar_Day) <= (extract(week(wednesday) from PARSE_DATE('%d/%m/%Y','21/10/2020')) + 1)
But this is not working for me.
Need help in resolving this. Thanks in Advance!
EXTRACT the week as the code already does. and the year as the weeks repeat every year.
GROUP BY the week and year. At this point I find it handy to make a STRUCT from the remaining fields as it simplifies the remaining code.
make another query that uses the query which did the GROUP BY, I used a WITH. In this last query, LEAD and LAG the data with a WINDOW by week.
Here's an example from a public dataset.
WITH
data_by_week AS (
SELECT
EXTRACT(year FROM date) AS year,
EXTRACT(week(wednesday) FROM date) AS week,
struct(
SUM(new_tested) as total_new_tested,
sum(new_recovered) as total_new_recovered
) as week_data
FROM
`bigquery-public-data.covid19_open_data.covid19_open_data`
GROUP BY
year,
week )
SELECT
year,
week,
LAG(week_data) OVER window_by_week AS previous_week,
week_data AS current_week,
LEAD(week_data) OVER window_by_week AS following_week
FROM
data_by_week
WINDOW
window_by_week AS ( ORDER BY year, week)
ORDER BY
year,
week
Can anyone help with calculating sales figures for month-until-date rolling sum on a data table carrying Datetime column and Sales figures column?
Using OVER in standard sql can help me calculate rows/dates preceding the current row, but I am having trouble with starting from day one of a month.
If you create columns for day, month, year (see:date extract function), you can use the month and year in the "PARTITION BY" part of your OVER function and the day in the "ORDER BY" part.
UPDATE
AliveToLearn worked it out: AVG(events_US) OVER (Partition by event_month, event_year ORDER BY day) AS moving_avg_month
I'm currently running the following sql statement in JasperReports Server to bring back my data using derived tables.
Select count(createddate) as ModulesCreatedDuringPastWeek,
count(updateddate) as ModulesUpdatedDuringPastWeek,
createddate,
updateddate
from merchendisingmodule
group by merchendisingmodule.createddate, merchendisingmodule.updateddate
However when grouping my data, I am only able to do it in Year, quarter, month and day. However for my report I'm needing the data to be group weeks, and so I was wondering what I will need to add to my code to do this.
DATEADD(D,-DATEPART(weekday,createddate)+1,createddate)
I use this method to prevent issues around the year transitions (week 53 in first days of januari and also in the last days of december, will group days together that are 360 days apart).
I use the first day of the week, instead of week numbers. I can use these dates to group by.
Also this will ensure that every week is 7 days long, instead of the last week of the year being only 3 or 4 days long.
Btw, in this example the first day of the week is sunday.
If your dates include time, use:
CAST(FLOOR(CAST(createddate AS FLOAT)) AS DATETIME)
instead of createddate in the above SYNTAX
I need to show distinct users per week. I have a date-visit column, and a user id, it is a big table with 1 billion rows.
I can change the date column from the CSVs to year,month, day columns. but how do I deduce the week from that in the query.
I can calculate the week from the CSV, but this is a big process step.
I also need to show how many distinct users visit day after day, looking for workaround as there is no date type.
any ideas?
To get the week of year number:
SELECT STRFTIME_UTC_USEC(TIMESTAMP('2015-5-19'), '%W')
20
If you have your date as a timestamp (i.e microseconds since the epoch) you can use the UTC_USEC_TO_DAY/UTC_USEC_TO_WEEK functions. Alternately, if you have an iso-formatted date string (e.g. "2012/03/13 19:00:06 -0700") you can call PARSE_UTC_USEC to turn the string into a timestamp and then use that to get the week or day.
To see an example, try:
SELECT LEFT((format_utc_usec(day)),10) as day, cnt
FROM (
SELECT day, count(*) as cnt
FROM (
SELECT UTC_USEC_TO_DAY(PARSE_UTC_USEC(created_at)) as day
FROM [publicdata:samples.github_timeline])
GROUP BY day
ORDER BY cnt DESC)
To show week, just change UTC_USEC_TO_DAY(...) to UTC_USEC_TO_WEEK(..., 0) (the 0 at the end is to indicate the week starts on Sunday). See the documentation for the above functions at https://developers.google.com/bigquery/docs/query-reference for more information.