update dataframe based on union function spark - scala

Hello guys i have a dataframe that needs to be updated based on another dataframe there are field that we are going to sum and other to just take the new value provided by the second dataframe , here what i did
val hist1 = spark.read
.format("csv")
.option("header", "true") //reading the headers
.load("C:/Users/MHT/Desktop/histocaisse_dte1.csv")
.withColumn("article_id", 'article_id.cast(LongType))
.withColumn("pos_id", 'pos_id.cast(LongType))
.withColumn("qte", 'qte.cast(LongType))
.withColumn("ca", 'ca.cast(DoubleType))
hist1.show
val hist2 = spark.read
.format("csv")
.option("header", "true") //reading the headers
.load("C:/Users/MHT/Desktop/his2.csv")
.withColumn("article_id", 'article_id.cast(LongType))
.withColumn("date", 'date.cast(DateType))
.withColumn("qte", 'qte.cast(LongType))
.withColumn("ca", 'ca.cast(DoubleType))
hist2.show
val df3 = hist1.unionAll(hist2)
//
val df4 = df3.groupBy("pos_id", "article_id").agg($"pos_id", $"article_id", max("date"), sum("qte"), sum("ca"))
df4.show
+------+----------+----------+---+----+----------+
|pos_id|article_id| date|qte| ca|sale_price|
+------+----------+----------+---+----+----------+
| 1| 1|2000-01-07| 3| 3.5| 14.3|
| 2| 2|2000-01-07| 15|12.0| 13.2|
| 3| 2|2000-01-07| 4| 1.2| 14.3|
| 4| 2|2000-01-07| 4| 1.2| 12.3|
+------+----------+----------+---+----+----------+
+------+----------+----------+---+----+----------+
|pos_id|article_id| date|qte| ca|sale_price|
+------+----------+----------+---+----+----------+
| 1| 1|2000-01-08| 3| 3.5| 14.5|
| 2| 2|2000-01-08| 15|12.0| 20.2|
| 3| 2|2000-01-08| 4| 1.2| 17.5|
| 4| 2|2000-01-08| 4| 1.2| 18.2|
| 5| 3|2000-01-08| 15| 1.2| 11.2|
| 6| 1|2000-01-08| 2|1.25| 13.5|
| 6| 2|2000-01-08| 2|1.25| 14.3|
+------+----------+----------+---+----+----------+
+------+----------+----------+--------+-------+
|pos_id|article_id| max(date)|sum(qte)|sum(ca)|
+------+----------+----------+--------+-------+
| 2| 2|2000-01-08| 30| 24.0|
| 3| 2|2000-01-08| 8| 2.4|
| 1| 1|2000-01-08| 6| 7.0|
| 5| 3|2000-01-08| 15| 1.2|
| 6| 1|2000-01-08| 2| 1.25|
| 6| 2|2000-01-08| 2| 1.25|
| 4| 2|2000-01-08| 8| 2.4|
+------+----------+----------+--------+-------+
How would the request be , if i want to append the field sale_price and consider the new sale_price provided by the second dataframe
how would this request be
val df4 = df3.groupBy("pos_id", "article_id").agg($"pos_id", $"article_id", max("date"), sum("qte"), sum("ca"))
Many thanks in advance

You can use join on the last line as the following
val df4 = df3.groupBy("pos_id", "article_id").agg(max("date"), sum("qte"), sum("ca")).join(hist2.select("pos_id", "article_id", "sale_price"), Seq("pos_id", "article_id"))
You should have your desired output

Related

Perform merge/insert on two spark dataframes with different schemas?

I have spark dataframe df and df1 both with different schemas.
DF:-
val DF = Seq(("1","acv","34","a","1"),("2","fbg","56","b","3"),("3","rty","78","c","5")).toDF("id","name","age","DBName","test")
+---+----+---+------+----+
| id|name|age|DBName|test|
+---+----+---+------+----+
| 1| acv| 34| a| 1|
| 2| fbg| 56| b| 3|
| 3| rty| 78| c| 5|
+---+----+---+------+----+
DF1:-
val DF1= Seq(("1","gbj","67","a","5"),("2","gbj","67","a","7"),("2","jku","88","b","8"),("4","jku","88","b",7"),("5","uuu","12","c","9")).toDF("id","name","age","DBName","col1")
+---+----+---+------+----+
| id|name|age|DBName|col1|
+---+----+---+------+----+
| 1| gbj| 67| a| 5|
| 2| gbj| 67| a| 7|
| 2| jku| 88| b| 8|
| 4| jku| 88| b| 7|
| 5| uuu| 12| c| 9|
+---+----+---+------+----+
I want to merge DF1 with DF based on value of id and DBName. So if my id and DBName already exists in DF then the record should be updated and if id and DBName doesn't exist then the new record should be added. So the resulting data frame should be like this:
+---+----+---+------+----+----+
| id|name|age|DBName|Test|col |
+---+----+---+------+----+----+
| 5| uuu| 12| c|NULL|9 |
| 2| jku| 88| b|NULL|8 |
| 4| jku| 88| b|NULL|7 |
| 1| gbj| 67| a|NULL|5 |
| 3| rty| 78| c|5 |NULL|
| 2| gbj| 67| a|NULL|7 |
+---+----+---+------+----+----+
I have tried so far
val updatedDF = DF.as("a").join(DF1.as("b"), $"a.id" === $"b.id" && $"a.DBName" === $"b.DBName", "outer").select(DF.columns.map(c => coalesce($"b.$c", $"b.$c") as c): _*)
Error:-
org.apache.spark.sql.AnalysisException: cannot resolve '`b.test`' given input columns: [b.DBName, a.DBName, a.name, b.age, a.id, a.age, b.id, a.test, b.name];;
You're selecting non-existent columns, and also there is a typo in the coalesce. You can follow the example below to fix your issue:
val updatedDF = DF.as("a").join(
DF1.as("b"),
$"a.id" === $"b.id" && $"a.DBName" === $"b.DBName",
"outer"
).select(
DF.columns.dropRight(1).map(c => coalesce($"b.$c", $"a.$c") as c)
:+ col(DF.columns.last)
:+ col(DF1.columns.last)
:_*
)
updatedDF.show
+---+----+---+------+----+----+
| id|name|age|DBName|test|col1|
+---+----+---+------+----+----+
| 5| uuu| 12| c|null| 9|
| 2| jku| 88| b| 3| 8|
| 4| jku| 88| b|null| 7|
| 1| gbj| 67| a| 1| 5|
| 3| rty| 78| c| 5|null|
| 2| gbj| 67| a|null| 7|
+---+----+---+------+----+----+

How to combine dataframes with no common columns?

I have 2 data frames
val df1 = Seq(("1","2","3"),("4","5","6")).toDF("A","B","C")
df1.show
+---+---+---+
| A| B| C|
+---+---+---+
| 1| 2| 3|
| 1| 2| 3|
+---+---+---+
and
val df2 = Seq(("11","22","33"),("44","55","66")).toDF("D","E","F")
df2.show
+---+---+---+
| D| E| F|
+---+---+---+
| 11| 22| 33|
| 44| 55| 66|
+---+---+---+
I need to combine the ones above to get
val df3 = Seq(("1","2","3","","",""),("4","5","6","","",""),("","","","11","22","33"),("","","","44","55","66"))
.toDF("A","B","C","D","E","F")
df3.show
+---+---+---+---+---+---+
| A| B| C| D| E| F|
+---+---+---+---+---+---+
| 1| 2| 3| | | |
| 4| 5| 6| | | |
| | | | 11| 22| 33|
| | | | 44| 55| 66|
+---+---+---+---+---+---+
Right now I'm creating the missing columns for all dataframes manually to get to a common structure and am then using a union. This code is specific to the dataframes and is not scalable
Looking for a solution that will work with x dataframes with y columns each
You can manually create missing columns in the two data frames and then union them:
import org.apache.spark.sql.DataFrame
val allCols = df1.columns.toSet.union(df2.columns.toSet).toArray
val createMissingCols = (df: DataFrame, allCols: Array[String]) => allCols.foldLeft(df)(
(_df, _col) => if (_df.columns.contains(_col)) _df else _df.withColumn(_col, lit(""))
).select(allCols.head, allCols.tail: _*)
// select is needed to make sure the two data frames have the same order of columns
createMissingCols(df1, allCols).union(createMissingCols(df2, allCols)).show
+---+---+---+---+---+---+
| E| F| A| B| C| D|
+---+---+---+---+---+---+
| | | 1| 2| 3| |
| | | 4| 5| 6| |
| 22| 33| | | | 11|
| 55| 66| | | | 44|
+---+---+---+---+---+---+
A much simpler way of doing this is creating a full outer join and setting the join expression/condition to false:
val df1 = Seq(("1","2","3"),("4","5","6")).toDF("A","B","C")
val df2 = Seq(("11","22","33"),("44","55","66")).toDF("D","E","F")
val joined = df1.join(df2, lit(false), "full")
joined.show()
+----+----+----+----+----+----+
| A| B| C| D| E| F|
+----+----+----+----+----+----+
| 1| 2| 3|null|null|null|
| 4| 5| 6|null|null|null|
|null|null|null| 11| 22| 33|
|null|null|null| 44| 55| 66|
+----+----+----+----+----+----+
if you then want to actually set the null values to empty string you can just add:
val withEmptyString = joined.na.fill("")
withEmptyString.show()
+---+---+---+---+---+---+
| A| B| C| D| E| F|
+---+---+---+---+---+---+
| 1| 2| 3| | | |
| 4| 5| 6| | | |
| | | | 11| 22| 33|
| | | | 44| 55| 66|
+---+---+---+---+---+---+
so in summary df1.join(df2, lit(false), "full").na.fill("") should do the trick.

Fill null or empty with next Row value with spark

Is there a way to replace null values in spark data frame with next row not null value. There is additional row_count column added for windows partitioning and ordering. More specifically, I'd like to achieve the following result:
+---------+-----------+ +---------+--------+
| row_count | id| |row_count | id|
+---------+-----------+ +------+-----------+
| 1| null| | 1| 109|
| 2| 109| | 2| 109|
| 3| null| | 3| 108|
| 4| null| | 4| 108|
| 5| 108| => | 5| 108|
| 6| null| | 6| 110|
| 7| 110| | 7| 110|
| 8| null| | 8| null|
| 9| null| | 9| null|
| 10| null| | 10| null|
+---------+-----------+ +---------+--------+
I tried with below code, It is not giving proper result.
val ss = dataframe.select($"*", sum(when(dataframe("id").isNull||dataframe("id") === "", 1).otherwise(0)).over(Window.orderBy($"row_count")) as "value")
val window1=Window.partitionBy($"value").orderBy("id").rowsBetween(0, Long.MaxValue)
val selectList=ss.withColumn("id_fill_from_below",last("id").over(window1)).drop($"row_count").drop($"value")
Here is a approach
Filter the non nulls (dfNonNulls)
Filter the nulls (dfNulls)
Find the right value for null id, using join and Window function
Fill the null dataframe (dfNullFills)
union dfNonNulls and dfNullFills
data.csv
row_count,id
1,
2,109
3,
4,
5,108
6,
7,110
8,
9,
10,
var df = spark.read.format("csv")
.option("header", "true")
.option("inferSchema", "true")
.load("data.csv")
var dfNulls = df.filter(
$"id".isNull
).withColumnRenamed(
"row_count","row_count_nulls"
).withColumnRenamed(
"id","id_nulls"
)
val dfNonNulls = df.filter(
$"id".isNotNull
).withColumnRenamed(
"row_count","row_count_values"
).withColumnRenamed(
"id","id_values"
)
dfNulls = dfNulls.join(
dfNonNulls, $"row_count_nulls" lt $"row_count_values","left"
).select(
$"id_nulls",$"id_values",$"row_count_nulls",$"row_count_values"
)
val window = Window.partitionBy("row_count_nulls").orderBy("row_count_values")
val dfNullFills = dfNulls.withColumn(
"rn", row_number.over(window)
).where($"rn" === 1).drop("rn").select(
$"row_count_nulls".alias("row_count"),$"id_values".alias("id"))
dfNullFills .union(dfNonNulls).orderBy($"row_count").show()
which results in
+---------+----+
|row_count| id|
+---------+----+
| 1| 109|
| 2| 109|
| 3| 108|
| 4| 108|
| 5| 108|
| 6| 110|
| 7| 110|
| 8|null|
| 9|null|
| 10|null|
+---------+----+

saving contents of df.show() as a string in pyspark

from pyspark.sql import SparkSession
sc = SparkSession.builder.getOrCreate()
prsn = sc.read.format("csv").option("delimiter", ",").option("header", "true").option("inferSchema", "true").load("C:/Users/amit.suar/IdeaProjects/LearningPyspark/prsn.csv")
prsn.show()
+-------------------------+------------------------+---+-----------+-----------------------------+
|PERSON_MEDIA_CONSUMER_KEY|PERSON_MEDIA_CONSUMER_ID|AGE|GENDER_CODE|EDUCATION_LEVEL_CATEGORY_CODE|
+-------------------------+------------------------+---+-----------+-----------------------------+
| 101| 3285854| 15| 1| 1|
| 102| 2313090| 25| 1| 3|
| 103| 2295854| 33| 2| 6|
| 104| 2295854| 33| 2| 6|
| 105| 2471554| 26| 2| 4|
| 106| 2471554| 26| 2| 4|
+-------------------------+------------------------+---+-----------+-----------------------------+
i want to capture this output as a string in a variable..how can i achieve it?
There is internal/private function that return the same string as .show() prints:
# Return dataframe as a table of first n records (20 by default)
dataframe._jdf.showString(n, 20)

In Spark Dataframe how to get duplicate records and distinct records in two dataframes?

I am working on a problem in which I am loading data from a hive table into spark dataframe and now I want all the unique accts in 1 dataframe and all duplicates in another. for example if I have acct id 1,1,2,3,4. I want to get 2,3,4 in one dataframe and 1,1 in another. How can I do this?
Depending on the version of spark you have, you could use window functions in datasets/sql like below:
Dataset<Row> New = df.withColumn("Duplicate", count("*").over( Window.partitionBy("id") ) );
Dataset<Row> Dups = New.filter(col("Duplicate").gt(1));
Dataset<Row> Uniques = New.filter(col("Duplicate").equalTo(1));
the above is written in java. should be similar in scala and read this on how to do in python.
https://databricks.com/blog/2015/07/15/introducing-window-functions-in-spark-sql.html
df.groupBy($"field1",$"field2"...).count.filter($"count" > 1).show()
val acctDF = List(("1", "Acc1"), ("1", "Acc1"), ("1", "Acc1"), ("2", "Acc2"), ("2", "Acc2"), ("3", "Acc3")).toDF("AcctId", "Details")
scala> acctDF.show()
+------+-------+
|AcctId|Details|
+------+-------+
| 1| Acc1|
| 1| Acc1|
| 1| Acc1|
| 2| Acc2|
| 2| Acc2|
| 3| Acc3|
+------+-------+
// Need to convert the DF to rdd to apply map and reduceByKey and again to DF to use it further more
val countsDF = acctDF.rdd.map(rec => (rec(0), 1)).reduceByKey(_+_).map(rec=> (rec._1.toString, rec._2)).toDF("AcctId", "AcctCount")
val accJoinedDF = acctDF.join(countsDF, acctDF("AcctId")===countsDF("AcctId"), "left_outer").select(acctDF("AcctId"), acctDF("Details"), countsDF("AcctCount"))
scala> accJoinedDF.show()
+------+-------+---------+
|AcctId|Details|AcctCount|
+------+-------+---------+
| 1| Acc1| 3|
| 1| Acc1| 3|
| 1| Acc1| 3|
| 2| Acc2| 2|
| 2| Acc2| 2|
| 3| Acc3| 1|
+------+-------+---------+
val distAcctDF = accJoinedDF.filter($"AcctCount"===1)
scala> distAcctDF.show()
+------+-------+---------+
|AcctId|Details|AcctCount|
+------+-------+---------+
| 3| Acc3| 1|
+------+-------+---------+
val duplAcctDF = accJoinedDF.filter($"AcctCount">1)
scala> duplAcctDF.show()
+------+-------+---------+
|AcctId|Details|AcctCount|
+------+-------+---------+
| 1| Acc1| 3|
| 1| Acc1| 3|
| 1| Acc1| 3|
| 2| Acc2| 2|
| 2| Acc2| 2|
+------+-------+---------+
(OR scala> duplAcctDF.distinct.show() )