cancel
Showing results forย 
Search instead forย 
Did you mean:ย 
Data Engineering
Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
cancel
Showing results forย 
Search instead forย 
Did you mean:ย 

Skip number of rows when reading CSV files

THIAM_HUATTAN
Valued Contributor

staticDataFrame = spark.read.format("csv")\ .option("header", "true").option("inferSchema", "true").load("/FileStore/tables/Consumption_2019/*.csv")

when above, I need an option to skip say first 4 lines on each CSV file, How do I do that?

8 REPLIES 8

mathan_pillai
Databricks Employee
Databricks Employee

Hi @THIAM HUAT TAN

I don't think there is a way to specify that when reading it. However, after reading it, you can create monotonically increasing id (new column), and then filter for those ids that are greater than 4.

Alternatively you can apply take(4) and create rdd out of it. Then apply subtract transformation between the original rdd and the small rdd.

please let us know whether it works for you

Thanks

THIAM_HUATTAN
Valued Contributor

databricks-data.png

My sample data is as above, and I need the data from Row 6 onwards, with Row 6 as the header. Row 1 to Row 5 are redundant. Not sure how to implement your suggestion. Thanks.

AnkitDwivedi
New Contributor II

I also have same issue. Is it resolved ?

What is the resolution ?

Please advise. Thanks

tony
New Contributor II

any resolution?

FabioKfouri
New Contributor II

I resolved it using the function monotonically_increasing_id and a little logic to set the Column Name.

To do this is necessary Java 1.8, because raise a error on function 'collect()' in Java 11.

df = df.withColumn('index', F.monotonically_increasing_id())
cols = df.columns
values = df.filter('index = 0').collect()  # here define the skipped lines
for i in range(len(cols)):
    if cols[i] != 'index':
        df = df.select(df.columns).withColumnRenamed(cols[i], values[0][i])

mstuder
New Contributor II

According to the docs of

spark.read.csv(...)
the
path
argument can be an RDD of strings:

path : str or list
     string, or list of strings, for input path(s), or RDD of Strings storing CSV rows.

With that, you may use

spark.sparkContext.textFile(...)
in combination with
zipWithIndex(...)
to perform the necessary row filtering. Putting things together this may look as follows:

n_skip_rows = ?
row_rdd = spark.sparkContext
    .textFile(your_csv_file) \
    .zipWithIndex() \
    .filter(lambda row: row[1] >= n_skip_rows) \
    .map(lambda row: row[0])
df = spark_session.read.csv(row_rdd, ...)

Hope that helps.

User16844409535
New Contributor II

You can provide the `skipRows` option while reading.

spark.read.format("csv").option("skipRows", 4).load("<filepath>")

Michael_Appiah
Contributor

The option... 

.option("skipRows", <number of rows to skip>) 

...works for me as well. However, I am surprised that the official Spark doc does not list it as a CSV Data Source Option: https://spark.apache.org/docs/latest/sql-data-sources-csv.html#data-source-option

@User16844409535 Did you find documentation on that somewhere else?

Connect with Databricks Users in Your Area

Join a Regional User Group to connect with local Databricks users. Events will be happening in your city, and you wonโ€™t want to miss the chance to attend and share knowledge.

If there isnโ€™t a group near you, start one and help create a community that brings people together.

Request a New Group