Spark Read CSV file into DataFrame

Spark SQL provides"path") to read a CSV file into Spark DataFrame and dataframe.write.csv("path") to save or write to the CSV file. Spark supports reading pipe, comma, tab, or any other delimiter/seperator files.

In this tutorial, you will learn how to read a single file, multiple files, all files from a local directory into DataFrame, and applying some transformations finally writing DataFrame back to CSV file using Scala.

Note: Spark out of the box supports to read files in CSV, JSON, TEXT, Parquet, and many more file formats into Spark DataFrame. 

Table of contents:

Spark Read CSV file into DataFrame

Using"path") or"csv").load("path") you can read a CSV file with fields delimited by pipe, comma, tab (and many more) into a Spark DataFrame, These methods take a file path to read from as an argument. You can find the zipcodes.csv at GitHub

val df ="src/main/resources/zipcodes.csv")

This example reads the data into DataFrame columns “_c0” for the first column and “_c1” for second and so on. and by default type of all these columns would be String.

 |-- _c0: string (nullable = true)
 |-- _c1: string (nullable = true)
 |-- _c2: string (nullable = true)

If you have a header with column names on file, you need to explicitly specify true for header option using option("header",true) not mentioning this, the API treats the header as a data record.

val df ="header",true)

It also reads all columns as a string (StringType) by default. I will explain in later sections how to read the schema (inferschema) from the header record and derive the column type based on the data.

When you use format("csv") method, you can also specify the Data sources by their fully qualified name (i.e., org.apache.spark.sql.csv), but for built-in sources, you can also use their short names (csv,jsonparquetjdbctext e.t.c). 

Read multiple CSV files

Using the method you can also read multiple CSV files, just pass all file names by separating comma as a path, for example : 

val df ="path1,path2,path3")

Read all CSV files in a directory

 We can read all CSV files from a directory into DataFrame just by passing the directory as a path to the csv() method.

val df ="Folder path")

Options while reading CSV file

Spark CSV dataset provides multiple options to work with CSV files. Below are some of the most important options explained with examples.


delimiter option is used to specify the column delimiter of the CSV file. By default, it is comma (,) character, but can be set to pipe (|), tab, space, or any character using this option.

val df2 ="delimiter"->","))


The default value set to this option is false when setting to true it automatically infers column types based on the data. Note that, it requires reading the data one more time to infer the schema.

val df2 ="inferSchema"->"true","delimiter"->","))

This option is used to read the first line of the CSV file as column names. By default the value of this option is false , and all column types are assumed to be a string.

val df2 ="inferSchema"->"true","delimiter"->",","header"->"true"))


When you have a column with a delimiter that used to split the columns, use quotes option to specify the quote character, by default it is ” and delimiters inside quotes are ignored. but using this option you can set any character.


Using nullValues option you can specify the string in a CSV to consider as null. For example, if you want to consider a date column with a value “1900-01-01” set null on DataFrame.


dateFormat option to used to set the format of the input DateType and TimestampType columns. Supports all java.text.SimpleDateFormat formats.

Note: Besides the above options, Spark CSV dataset also supports many other options, please refer to this article for details.

Reading CSV files with a user-specified custom schema

If you know the schema of the file ahead and do not want to use the inferSchema option for column names and types, use user-defined custom column names and type using schema option.

val schema = new StructType()
val df_with_schema ="csv")
      .option("header", "true")

Applying DataFrame transformations

Once you have created DataFrame from the CSV file, you can apply all transformation and actions DataFrame support. Please refer to the link for more details. 

Write Spark DataFrame to CSV file

Use the write() method of the Spark DataFrameWriter object to write Spark DataFrame to a CSV file. For detailed example refer to Writing Spark DataFrame to CSV File using Options.



While writing a CSV file you can use several options. for example, header to output the DataFrame column names as header record and delimiter to specify the delimiter on the CSV output file.


Other options available quote,escape,nullValue,dateFormat,quoteMode .

Saving modes

Spark DataFrameWriter also has a method mode() to specify SaveMode; the argument to this method either takes below string or a constant from SaveMode class.

overwrite – mode is used to overwrite the existing file, alternatively, you can use SaveMode.Overwrite.

append – To add the data to the existing file, alternatively, you can use SaveMode.Append.

ignore – Ignores write operation when the file already exists, alternatively you can use SaveMode.Ignore.

errorifexists or error – This is a default option when the file already exists, it returns an error, alternatively, you can use SaveMode.ErrorIfExists.



In this tutorial, you have learned how to read a CSV file, multiple csv files and all files from a local folder into Spark DataFrame, using multiple options to change the default behavior and write CSV files back to DataFrame using different save options.


Happy Learning !!

Naveen (NNK)

I am Naveen (NNK) working as a Principal Engineer. I am a seasoned Apache Spark Engineer with a passion for harnessing the power of big data and distributed computing to drive innovation and deliver data-driven insights. I love to design, optimize, and managing Apache Spark-based solutions that transform raw data into actionable intelligence. I am also passion about sharing my knowledge in Apache Spark, Hive, PySpark, R etc.

Leave a Reply

This Post Has 23 Comments

  1. Divyesh

    Hello NNK,

    Huge fan of the website. I was trying to read multiple csv files located in different folders as:[“path_1″,”path_2″,”path_3”], header = True)

    and was successfully able to do that. However, when running the program from spark-submit says that spark module not found.

    Can you help.

    1. NNK

      Thanks Divyesh for your comments. Could you please share your complete stack trace error? If you have already resolved the issue, please comment here, others would get benefit from your solution.

      Thanks again. Happy Learning !!

  2. Dhineshkumar R

    Hi Team,

    Actually headers in my csv file starts from 3rd row? How can I configure in such cases?
    Your help is highly appreciated.

    1. NNK

      Hi Dhinesh, By default Spark-CSV can’t handle it, however, you can do it by custom code as mentioned below.

      1) Read the CSV file using spark-csv as if there is no header
      2) use filter on DataFrame to filter out header row
      3) used the header row to define the columns of the DataFrame
      4) finally assign the columns to DataFrame

      Hope it gives you some idea.

  3. Ashwin s

    Hi NNK,
    I’m getting an error while trying to read a csv file from github using above mentioned process.
    “Py4JJavaError: An error occurred while calling o100.csv.
    : No FileSystem for scheme: ……..”.
    Kindly help.Thanks in Advance.

  4. Suraj Nepram

    all the column values are coming as null when csv is read with schema
    val df_with_schema =“csv”)
    .option(“header”, “true”)

    How do I fix this? I am using a window system. reading the csv without schema works fine

  5. Boris

    Thank you for the information and explanation!

    1. NNK

      Thanks Boris for reading.

  6. Syahirah

    i get it can read multiple files, but may i know if the CSV files have the same attributes/column or not? is it possible to have multiple files such as CSV1 is personal data, CSV2 is the call usage, CSV3 is the data usage and combined it together to put in dataframe.

    1. NNK

      When you reading multiple CSV files from a folder, all CSV files should have the same attributes and columns. You can’t read different CSV files into the same DataFrame.

  7. Anonymous

    hi there. I did the schema and got the appropriate types bu i cannot use the describe function.

    1. NNK

      May I know where are you using the describe function?

  8. Anonymous

    Hi, nice article! I am wondering how to read from CSV file which has more than 22 columns and create a data frame using this data

  9. Bindhu


    I want to rename a part of file name in a folder. example: XXX_07_08 to XXX_0700008. Please guide

    1. NNK

      In order to rename file name you have to use hadoop file system API

  10. K.T. Wong


    Great website, and extremely helpfull. My appreciation and gratitude 😉

    I want to ingest data from a folder containing csv files, but upon ingestion I want one column containing the filename of the data that is being ingested. Any ideas on how to accomplish this?

    K.T. Wong

    1. NNK

      Hi Wong, Thanks for your kind words.
      After reading a CSV file into DataFrame use the below statement to add a new column. please comment if this works.

      import org.apache.spark.sql.functions.lit
      df.withColumn(“fileName”, lit(“file-name”)).

  11. jsp

    Where can i find the data files like zipcodes.csv

  12. Mark

    Nice content but adverts make it unreadable

You are currently viewing Spark Read CSV file into DataFrame
Photo by Lewis Ngugi on Unsplash