[英]BigQuery - concatenate all rows of column until particular value reached
我想連接特定列的所有行,直到達到某個值。
因此,在我的示例中,我想使用“>”作為分隔符連接“PageType”列,直到到達字符串值為“Checkout”的行
它應該看起來像這樣: Home > PLP > PDP > PLP > PDP > PLP > PDP > Checkout
先感謝您。
基本上,在聚合之前進行過濾:
select visitid,
string_agg(pagetype, '>' order by time) as checkout_path
from (select t.*,
min(case when pagetype = 'Checkout' then time end) over (partition by visitid) as checkout_time
from t
) t
where time <= checkout_time
group by visitid;
以下是 BigQuery 標准 SQL
#standardSQL
SELECT visitId,
SPLIT(path, ' > Checkout')[OFFSET(0)] || ' > Checkout' path
FROM (
SELECT visitId,
STRING_AGG(PageType, ' > ' ORDER BY time) AS path
FROM `project.dataset.table`
GROUP BY visitId
)
WHERE path LIKE '% > Checkout'
當應用於您的問題中的示例數據時 - output 是
Row visitId path
1 1597865846 Home > PLP > PDP > PLP > PDP > PLP > PDP > Checkout
我看到您使用了 Google Analytics 數據,所以我使用了公開可用的示例數據。 我寫了一個子查詢(包含另一個停止條件的子查詢),而不是不必要的橫向交叉連接。
DECLARE stopPage DEFAULT '/basket.html';
SELECT
fullvisitorid,
visitstarttime,
(SELECT STRING_AGG(page.pagePath, ' > ' ORDER BY time ASC) FROM UNNEST(hits)
WHERE type = 'PAGE'
AND time <= (select min(time) from unnest(hits) where page.pagePath=stopPage)) AS path
#ARRAY(SELECT AS STRUCT ecommerceAction, page FROM UNNEST(hits) WHERE type = 'PAGE' ) AS hits
FROM `bigquery-public-data.google_analytics_sample.ga_sessions_20170801`
WHERE exists(SELECT 1 FROM unnest(hits) WHERE page.pagePath=stopPage)
LIMIT 1000
由於沒有頁面類型,我改用頁面路徑(我隨機選擇了 /basket.html)- 我認為您可以輕松調整條件。
您可以運行此查詢以查看結果。
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.