将 csv 文件读取为 dataframe 时模式错误

Question

trying to read a csv file into a dataframe simple code

df = spark.read.csv("1.csv")

i got

    df.printSchema()
root
 |-- _c0: string (nullable = true)

also i try this

db = spark.read.csv("1.csv", header=True, inferSchema= "True")
db.printSchema()
root
 |--                   id                  |                      date                      |                              cases                               |                      country                      |                       deaths                       |   cities   |    per_cap     | 

Thanks in advance for your help

Answer 1

apparently, your line seperator is a pipe | .

try:

db = spark.read.csv("1.csv", sep='|', header=True, inferSchema= "True")

for col in db.columns:
    db = db.withColumnRenamed(col, col.strip())

Answer 2

you should create your own schema.
using scala:

val schemaExpected = new StructType()
.add("id" , StringType , nullable = true)
.add("date" , DateType, nullable = true)
...
.add(deaths , IntegerType, nullable = true)

then you can read your DataFrame:

val db = spark.read.option("header","true").schema(schemaExpected).csv("1.csv")