In this article, I will show you how to combine two Spark DataFrames that have no common columns.
Table of Contents
For example, if we have the two following DataFrames:
val df1 = Seq(
("001","002","003"),
("004","005","006")
).toDF("A","B","C")
val df2 = Seq(
("011","022","033"),
("044","055","066")
).toDF("D","E","F")
The output I want to get looks like this:
+----+----+----+----+----+----+
| A| B| C| D| E| F|
+----+----+----+----+----+----+
| 001| 002| 003|null|null|null|
| 004| 005| 006|null|null|null|
|null|null|null| 011| 022| 033|
|null|null|null| 044| 055| 066|
+----+----+----+----+----+----+
This can be easily achieved by using the full outer join with the condition set to false:
df1.join(df2, lit(false), "full")
It works because the full outer join takes all rows from both DataFrames, so we end up with all rows, and we use lit(false) as the joining condition, which ensures that there will be no matches between both DataFrames.
Get Weekly AI Implementation Insights
Join engineering leaders who receive my analysis of common AI production failures and how to prevent them. No fluff, just actionable techniques.