Duplicating the contents of a string column a specified number of times

PySpark @ Freshers.in

The repeat function in PySpark is used to duplicate the contents of a string column a specified number of times. It’s a part of the pyspark.sql.functions module and is particularly useful in data augmentation and formatting tasks. This article aims to elucidate the functionality of repeat in PySpark, illustrated with practical examples. The repeat function in PySpark is a handy tool for various data manipulation tasks, especially in scenarios requiring data amplification or specific formatting. Its simplicity belies its potential to greatly enhance the versatility of data handling in PySpark.


repeat(col, n)

col: The string column to be repeated.
n: The number of times to repeat the column content.

Example: Data augmentation

Consider a scenario where we want to repeat the names in a dataset a certain number of times for data augmentation purposes.

Dataset Example:


Let’s say we want to repeat each name 3 times.

Step-by-Step Implementation:

Initializing PySpark:

Set up your PySpark session and import the necessary function.

from pyspark.sql import SparkSession
from pyspark.sql.functions import repeat
spark = SparkSession.builder.appName("repeat_function_example").getOrCreate()

Creating the dataframe:

Create a DataFrame with the provided names.

data = [("Sachin",), ("Ram",), ("Raju",), ("David",), ("Wilson",)]
df = spark.createDataFrame(data, ["Name"])

Applying repeat:

Use the repeat function to repeat each name 3 times.

repeated_df = df.withColumn("Repeated_Name", repeat(df.Name, 3))


Name Repeated_Name
Sachin SachinSachinSachin
Ram RamRamRam
Raju RajuRajuRaju
David DavidDavidDavid
Wilson WilsonWilsonWilson

Use case/scenario : NLP and Text data augmentation

In many NLP tasks, such as sentiment analysis, text classification, or language modeling, having a large and varied dataset is crucial for training robust machine learning models. However, often, datasets may be imbalanced or lack sufficient examples of certain classes or patterns.

Application of repeat:

The repeat function can be employed to artificially augment text data, especially to emphasize certain words or phrases that are critical for the model to learn. For example, in a sentiment analysis task, certain keywords that strongly indicate a sentiment (like “great” for positive or “terrible” for negative) can be repeated to amplify their presence in the training data. This can help in scenarios where such keywords are underrepresented.


Suppose you have a dataset of customer reviews, and you are building a model to classify them into ‘Positive’ or ‘Negative’. You might find that the word “excellent” is a strong indicator of a positive review but is not frequently used in your dataset.

Original dataset:

Review Sentiment
The service was excellent Positive
An excellent experience Positive

Augmented dataset using repeat:

Review Sentiment
The service was excellent excellent Positive
An excellent excellent experience Positive

By repeating the word “excellent”, the model gets more examples where this word is a key feature, potentially improving its ability to recognize positive sentiments associated with it.


  • Enhances Feature Representation: Helps in emphasizing certain words or phrases, making them more prominent features for model training.
  • Improves Model Robustness: Especially useful in scenarios with limited or imbalanced data, aiding in building more robust NLP models.
  • Flexible and Easy to Implement: The simplicity of the repeat function allows for easy integration into data preprocessing pipelines.


  • Risk of Overfitting: Overusing this technique might lead to models that are too focused on the repeated features and may not generalize well.
  • Balance and Diversity: It’s important to ensure that the augmentation does not bias the model excessively towards certain features or classes.

Spark important urls to refer

  1. Spark Examples
  2. PySpark Blogs
  3. Bigdata Blogs
  4. Spark Interview Questions
  5. Official Page
Author: user