当前位置：网站首页>Scala104-Spark.sql的内置日期时间函数
Scala104-Spark.sql的内置日期时间函数
2022-08-04 18:29:00 【51CTO】
有些时候我们会直接用df.createOrReplaceTempView(temp)创建临时表，用sql去计算。sparkSQL有些语法和hql不一样，做个笔记。
- <scala.version>2.11.12</scala.version>
- <spark.version>2.4.3</spark.version>
val
builder
=
SparkSession
.
builder()
.
appName(
"learningScala")
.
config(
"spark.executor.heartbeatInterval",
"60s")
.
config(
"spark.network.timeout",
"120s")
.
config(
"spark.serializer",
"org.apache.spark.serializer.KryoSerializer")
.
config(
"spark.kryoserializer.buffer.max",
"512m")
.
config(
"spark.dynamicAllocation.enabled",
false)
.
config(
"spark.sql.inMemoryColumnarStorage.compressed",
true)
.
config(
"spark.sql.inMemoryColumnarStorage.batchSize",
10000)
.
config(
"spark.sql.broadcastTimeout",
600)
.
config(
"spark.sql.autoBroadcastJoinThreshold",
-
1)
.
config(
"spark.sql.crossJoin.enabled",
true)
.
master(
"local[*]")
val
spark
=
builder.
getOrCreate()
spark.
sparkContext.
setLogLevel(
"ERROR")
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
builder: org.apache.spark.sql.SparkSession.Builder = [email protected]
spark: org.apache.spark.sql.SparkSession = [email protected]
- 1.
- 2.
var
df1
=
Seq(
(
1,
"2019-04-01 11:45:50",
11.15,
"2019-04-02 11:45:49"),
(
2,
"2019-05-02 11:56:50",
10.37,
"2019-05-02 11:56:51"),
(
3,
"2019-07-21 12:45:50",
12.11,
"2019-08-21 12:45:50"),
(
4,
"2019-08-01 12:40:50",
14.50,
"2020-08-03 12:40:50"),
(
5,
"2019-01-06 10:00:50",
16.39,
"2019-01-05 10:00:50")
).
toDF(
"id",
"startTimeStr",
"payamount",
"endTimeStr")
df1
=
df1.
withColumn(
"startTime",
$
"startTimeStr".
cast(
"Timestamp"))
.
withColumn(
"endTime",
$
"endTimeStr".
cast(
"Timestamp"))
df1.
printSchema
df1.
show()
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
root
|-- id: integer (nullable = false)
|-- startTimeStr: string (nullable = true)
|-- payamount: double (nullable = false)
|-- endTimeStr: string (nullable = true)
|-- startTime: timestamp (nullable = true)
|-- endTime: timestamp (nullable = true)
+---+-------------------+---------+-------------------+-------------------+-------------------+
| id| startTimeStr|payamount| endTimeStr| startTime| endTime|
+---+-------------------+---------+-------------------+-------------------+-------------------+
| 1|2019-04-01 11:45:50| 11.15|2019-04-02 11:45:49|2019-04-01 11:45:50|2019-04-02 11:45:49|
| 2|2019-05-02 11:56:50| 10.37|2019-05-02 11:56:51|2019-05-02 11:56:50|2019-05-02 11:56:51|
| 3|2019-07-21 12:45:50| 12.11|2019-08-21 12:45:50|2019-07-21 12:45:50|2019-08-21 12:45:50|
| 4|2019-08-01 12:40:50| 14.5|2020-08-03 12:40:50|2019-08-01 12:40:50|2020-08-03 12:40:50|
| 5|2019-01-06 10:00:50| 16.39|2019-01-05 10:00:50|2019-01-06 10:00:50|2019-01-05 10:00:50|
+---+-------------------+---------+-------------------+-------------------+-------------------+
df1: org.apache.spark.sql.DataFrame = [id: int, startTimeStr: string ... 4 more fields]
df1: org.apache.spark.sql.DataFrame = [id: int, startTimeStr: string ... 4 more fields]
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
timestamp转string
把timestamp转换成对应格式字符串
- date_format把timestamp转换成对应的字符串
- 字符串格式用"yyyyMMdd"表示
root
|-- yyyyMMdd: string (nullable = true)
|-- yyyy_MM_dd: string (nullable = true)
|-- yyyy: string (nullable = true)
+--------+----------+----+
|yyyyMMdd|yyyy_MM_dd|yyyy|
+--------+----------+----+
|20190401|2019-04-01|2019|
|20190502|2019-05-02|2019|
|20190721|2019-07-21|2019|
|20190801|2019-08-01|2019|
|20190106|2019-01-06|2019|
+--------+----------+----+
sql: String =
"
SELECT date_format(startTime,'yyyyMMdd') AS yyyyMMdd,
date_format(startTime,'yyyy-MM-dd') AS yyyy_MM_dd,
date_format(startTime,'yyyy') AS yyyy
FROM TEMP
"
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
timestamp转date
- to_date可以把timestamp转换成date类型
root
|-- startTime: timestamp (nullable = true)
|-- endTime: timestamp (nullable = true)
|-- startDate: date (nullable = true)
|-- endDate: date (nullable = true)
+-------------------+-------------------+----------+----------+
| startTime| endTime| startDate| endDate|
+-------------------+-------------------+----------+----------+
|2019-04-01 11:45:50|2019-04-02 11:45:49|2019-04-01|2019-04-02|
|2019-05-02 11:56:50|2019-05-02 11:56:51|2019-05-02|2019-05-02|
|2019-07-21 12:45:50|2019-08-21 12:45:50|2019-07-21|2019-08-21|
|2019-08-01 12:40:50|2020-08-03 12:40:50|2019-08-01|2020-08-03|
|2019-01-06 10:00:50|2019-01-05 10:00:50|2019-01-06|2019-01-05|
+-------------------+-------------------+----------+----------+
sql: String =
SELECT startTime,endTime,
to_date(startTime) AS startDate,
to_date(endTime) AS endDate
FROM TEMP
df2: org.apache.spark.sql.DataFrame = [startTime: timestamp, endTime: timestamp ... 2 more fields]
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
求时间差
- 天数差函数datediff可以应用在timestamp中，也可应用在date类型中，单位是自然天，而不是24小时
- 月份差函数months_between同样可以，月度的单位好像是不固定的，即31天or30天
df2.
createOrReplaceTempView(
"temp")
var
sql
=
"""
SELECT startTime,
endTime,
datediff(endTime,startTime) AS dayInterval1,
datediff(endDate,startDate) AS dayInterval2,
months_between(endTime,startTime) AS monthInterval1,
months_between(endDate,startDate) AS monthInterval2
FROM TEMP
"""
// spark.sql(sql).printSchema
spark.
sql(
sql).
show()
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
+-------------------+-------------------+------------+------------+--------------+--------------+
| startTime| endTime|dayInterval1|dayInterval2|monthInterval1|monthInterval2|
+-------------------+-------------------+------------+------------+--------------+--------------+
|2019-04-01 11:45:50|2019-04-02 11:45:49| 1| 1| 0.03225769| 0.03225806|
|2019-05-02 11:56:50|2019-05-02 11:56:51| 0| 0| 0.0| 0.0|
|2019-07-21 12:45:50|2019-08-21 12:45:50| 31| 31| 1.0| 1.0|
|2019-08-01 12:40:50|2020-08-03 12:40:50| 368| 368| 12.06451613| 12.06451613|
|2019-01-06 10:00:50|2019-01-05 10:00:50| -1| -1| -0.03225806| -0.03225806|
+-------------------+-------------------+------------+------------+--------------+--------------+
sql: String =
"
SELECT startTime,
endTime,
datediff(endTime,startTime) AS dayInterval1,
datediff(endDate,startDate) AS dayInterval2,
months_between(endTime,startTime) AS monthInterval1,
months_between(endDate,startDate) AS monthInterval2
FROM TEMP
"
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
Ref
2020-03-24 于南京市江宁区九龙湖
边栏推荐
- npm配置国内镜像（淘宝镜像）
- 电源测试系统-ATE电源测试系统-ACDC电源模块测试系统NSAT-8000
- powershell和cmd对比
- golang安装和基础配置
- ATF中断处理的设计模型
- 用Excel绘制统计图
- Flask framework implementations registered encryption, a Flask enterprise class learning 】 【
- 解决错误：The package-lock.json file was created with an old version of npm
- mysql cdc 为什么需要RELOAD 这个权限？这个权限在采集数据的过程中的作用是什么？有哪
- 在表格数据集上训练变分自编码器 (VAE)示例
猜你喜欢
随机推荐
MySQL安装教程（详细）
powershell和cmd对比
关于使用腾讯云HiFlow场景连接器每天提醒签到打卡
如何封装 svg
使用.NET简单实现一个Redis的高性能克隆版（二）
Investigation and Research Based on the Involution Behavior of College Students
悦刻难回巅峰
Enterprise survey correlation analysis case
mysql cdc 为什么需要RELOAD 这个权限？这个权限在采集数据的过程中的作用是什么？有哪
PHP代码审计7—文件上传漏洞
老电脑怎么重装系统win10
如何模拟后台API调用场景，很细！
Babbitt | Metaverse daily must-read: Weibo animation will recruit all kinds of virtual idols around the world and provide support for them...
阿里云技术专家秦隆：云上如何进行混沌工程？
ATF中断处理的设计模型
clickhouse online and offline table
The Industrial Metaverse Brings Changes to Industry
Flask框架实现注册加密功能详解【Flask企业课学习】
【web自动化测试】playwright安装失败怎么办
Win10只读文件夹怎么删除