AnalysisException: u"cannot resolve 'name' given input columns: [ list] in sqlContext in spark

Question

I tried a simple example like:

data = sqlContext.read.format("csv").option("header", "true").option("inferSchema", "true").load("/databricks-datasets/samples/population-vs-price/data_geo.csv")

data.cache() # Cache data for faster reuse
data = data.dropna() # drop rows with missing values
data = data.select("2014 Population estimate", "2015 median sales price").map(lambda r: LabeledPoint(r[1], [r[0]])).toDF()

It works well, But when i try something very similar like:

data = sqlContext.read.format("csv").option("header", "true").option("inferSchema", "true").load('/mnt/%s/OnlineNewsTrainingAndValidation.csv' % MOUNT_NAME)

data.cache() # Cache data for faster reuse
data = data.dropna() # drop rows with missing values
data = data.select("timedelta", "shares").map(lambda r: LabeledPoint(r[1], [r[0]])).toDF()
display(data)

It raise error: AnalysisException: u"cannot resolve 'timedelta' given input columns: [ data_channel_is_tech,...

off-course I imported LabeledPoint and LinearRegression

What could be wrong?

Even the simpler case

df_cleaned = df_cleaned.select("shares")

raises same AnalysisException (error).

*please note: df_cleaned.printSchema() works well.

Answer 1

I found the issue: some of the column names contain white spaces before the name itself. So

data = data.select(" timedelta", " shares").map(lambda r: LabeledPoint(r[1], [r[0]])).toDF()

worked. I could catch the white spaces using

assert " " not in ''.join(df.columns)  

Now I am thinking of a way to remove the white spaces. Any idea is much appreciated!

Answer 2

Because header contains spaces or tabs,remove spaces or tabs and try

1) My example script

from pyspark.sql import SparkSession
spark = SparkSession \
.builder \
.appName("Python Spark SQL basic example") \
.config("spark.some.config.option", "some-value") \
.getOrCreate()

df=spark.read.csv(r'test.csv',header=True,sep='^')
print("#################################################################")
print df.printSchema()
df.createOrReplaceTempView("test")
re=spark.sql("select max_seq from test")
print(re.show())
print("################################################################")

2) Input file,here 'max_seq ' contains space so we are getting bellow exception

Trx_ID^max_seq ^Trx_Type^Trx_Record_Type^Trx_Date

Traceback (most recent call last):
  File "D:/spark-2.1.0-bin-hadoop2.7/bin/test.py", line 14, in <module>
    re=spark.sql("select max_seq from test")
  File "D:\spark-2.1.0-bin-hadoop2.7\python\lib\pyspark.zip\pyspark\sql\session.py", line 541, in sql
  File "D:\spark-2.1.0-bin-hadoop2.7\python\lib\py4j-0.10.4-src.zip\py4j\java_gateway.py", line 1133, in __call__
  File "D:\spark-2.1.0-bin-hadoop2.7\python\lib\pyspark.zip\pyspark\sql\utils.py", line 69, in deco
pyspark.sql.utils.AnalysisException: u"cannot resolve '`max_seq`' given input columns: [Venue_City_Name, Trx_Type, Trx_Booking_Status_Committed, Payment_Reference1, Trx_Date, max_seq , Event_ItemVariable_Name, Amount_CurrentPrice, cinema_screen_count, Payment_IsMyPayment, r

2) Remove space after 'max_seq' column then it will work fine

Trx_ID^max_seq^Trx_Type^Trx_Record_Type^Trx_Date


17/03/20 12:16:25 INFO DAGScheduler: Job 3 finished: showString at <unknown>:0, took 0.047602 s
17/03/20 12:16:25 INFO CodeGenerator: Code generated in 8.494073 ms
  max_seq
    10
    23
    22
    22
only showing top 20 rows

None
##############################################################

Answer 3

As there were tabs in my input file, removing the tabs or spaces in the header helped display the answer.

My example:

saledf = spark.read.csv("SalesLTProduct.txt", header=True, inferSchema= True, sep='\t')


saledf.printSchema()

root
|-- ProductID: string (nullable = true)
|-- Name: string (nullable = true)
|-- ProductNumber: string (nullable = true)

saledf.describe('ProductNumber').show()

 +-------+-------------+
 |summary|ProductNumber|
 +-------+-------------+
 |  count|          295|
 |   mean|         null|
 | stddev|         null|
 |    min|      BB-7421|
 |    max|      WB-H098|
 +-------+-------------+

Answer 4

If you don't have whitespaces in headers, this error also raised when you not specify headers for csv at all like this:

df = sqlContext.read.csv('data.csv')

So you need to change it to this:

df = sqlContext.read.csv('data.csv', header=True)

Answer 5

Recently, I came across this issue while working on Azure synapse analytics; my error was the same.

analysisexception: cannot resolve '`xxxxxx`' given input columns: [];; 'filter ('passenger_count > 0) +- relation[] csv traceback (most recent call last):

 file "/opt/spark/python/lib/pyspark.zip/pyspark/sql/dataframe.py", line 1364, in filter jdf = self._jdf.filter(condition._jc) file "/opt/spark/python/lib/py4j-0.10.7-src.zip/py4j/java_gateway.py", line 1257, in __call__ answer, self.gateway_client, self.target_id, self.name)
 file "/opt/spark/python/lib/pyspark.zip/pyspark/sql/utils.py", line 75, in deco raise analysisexception(s.split(': ', 1)[1], stacktrace)""

This error came due to improper wording in our code or in CSV file use this code to read csv file:

-df = spark.read.load("examples/src/main/resources/people.csv",
                     format="csv", sep=";", inferSchema="true", header="true")

If you are again stuck somewhere in synapse or pyspark visit this site FOR Error info: https://docs.actian.com/avalanche/index.html#page/User/Common_Data_Loading_Error_Messages.htm

and for more info visit documentation: https://spark.apache.org/docs/latest/api/python/

Answer 6

I had a similar issue while trying to do max aggregation on a dataframe. It occurred I didn't import pyspark.sql.functions.max, but used built-in Python's max() function and that was causing the error.