【Pyspark】 一列变多列 、分割 一行中的list分割转为多列 explode，多列变一列

【Pyspark】 一列变多列   

分割 一行中的list分割转为多列 explode

官方例子：Python pyspark.sql.functions.explode() Examples

https://www.programcreek.com/python/example/98237/pyspark.sql.functions.explode

 

根据某个字段内容进行分割，然后生成多行，这时可以使用explode方法

Eg:

df.explode("c3","c3_"){time: String => time.split(" ")}.show(False)

https://blog.csdn.net/anshuai_aw1/article/details/87881079#4.4%C2%A0%E5%88%86%E5%89%B2%EF%BC%9A%E8%A1%8C%E8%BD%AC%E5%88%97

 

Eg:

from pyspark.sql import Row
eDF = sqlContext.createDataFrame([Row(a=1, intlist=[1,2,3], mapfield={"a": "b"})])
eDF.select(explode(eDF.intlist).alias("anInt")).collect()
Out: [Row(anInt=1), Row(anInt=2), Row(anInt=3)]

 

来自 <http://spark.apache.org/docs/1.6.0/api/python/pyspark.sql.html?highlight=except#pyspark.sql.functions.exp>

 

Eg:

from pyspark.sql import Row

from pyspark.sql.functions import explode

eDF = spark.createDataFrame([Row(a=1, intlist=[1,2,3], mapfield={"a": "b"})])



eDF.show()

eDF.select(explode(eDF.intlist).alias("anInt")).show()

 

Eg:

df2=df1.select(explode(df1.line).alias("line_new_name")) 

 

 

eg：

### 将结果分割，并分裂为不同列

### 将结果分割，并分裂为不同列
df = df.withColumn('res', getLabel(df['data'])).withColumn("res2", explode(col('res')))   #getLabel为udf函数
df_split = df.withColumn("s", split(df['res2'], ",")) 
df_split = df_split.withColumn('id', df_split["s"].getItem(0)) 
df_split = df_split.withColumn('label', df_split["s"].getItem(1))

 

 

多列变一列

A . 拼接:

 

concat_ws(',',collect_list(concat(concat(concat(concat(rank,'@@'),lon),'@@'),lat))) as info

 

B .collect_list、collect_set

返回list，

collect_set去除重复元素；collect_list不去除重复元素

 

官方例子：

https://www.programcreek.com/python/example/98235/pyspark.sql.functions.collect_list

 

pyspark.sql.functions.collect_list(col) 不去重

Aggregate function: returns a list of objects with duplicates.

New in version 1.6.

pyspark.sql.functions.collect_set(col)去重

Aggregate function: returns a set of objects with duplicate elements eliminated

 

但是collect_list不能保持原始的顺序,需要添加rank辅助列

拼接之后在使用的时候再split，不然直接split的话顺序就又乱了

eg: 

# 按rank排序后拼接
df_out =spark.sql('''
select cityid, 
    regexp_replace(
       concat_ws(',', 
             sort_array(
                    collect_list(  
                     concat_ws(  ':',
                                 lpad(cast(rank as string),2,'0'), 
                                 concat_ws("@@",concat_ws("@@", concat_ws("@@", concat_ws("@@", concat_ws("@@",poiid,poiname) ,lon),lat) ,hour),rank) 
                    )
                    )
              )  
        )
     , '\\[0-9]+\:',
     '')   
        
as line_string 
from  db.tablename where rank<100 group by id ''').distinct()
 

 

来源：CSDN

作者：sunflower_sara

链接：https://blog.csdn.net/sunflower_sara/article/details/104044252