Data sources

In this section, we introduce how to use data source in ML to load data. Besides some general data sources such as Parquet, CSV, JSON and JDBC, we also provide some specific data sources for ML.

Table of Contents

Image data source

This image data source is used to load image files from a directory, it can load compressed image (jpeg, png, etc.) into raw image representation via ImageIO in Java library. The loaded DataFrame has one StructType column: “image”, containing image data stored as image schema. The schema of the image column is:

ImageDataSource implements a Spark SQL data source API for loading image data as a DataFrame.

scala> val df = spark.read.format("image").option("dropInvalid", true).load("data/mllib/images/origin/kittens")
df: org.apache.spark.sql.DataFrame = [image: struct<origin: string, height: int ... 4 more fields>]

scala> df.select("image.origin", "image.width", "image.height").show(truncate=false)
+-----------------------------------------------------------------------+-----+------+
|origin                                                                 |width|height|
+-----------------------------------------------------------------------+-----+------+
|file:///spark/data/mllib/images/origin/kittens/54893.jpg               |300  |311   |
|file:///spark/data/mllib/images/origin/kittens/DP802813.jpg            |199  |313   |
|file:///spark/data/mllib/images/origin/kittens/29.5.a_b_EGDP022204.jpg |300  |200   |
|file:///spark/data/mllib/images/origin/kittens/DP153539.jpg            |300  |296   |
+-----------------------------------------------------------------------+-----+------+

ImageDataSource implements Spark SQL data source API for loading image data as a DataFrame.

Dataset<Row> imagesDF = spark.read().format("image").option("dropInvalid", true).load("data/mllib/images/origin/kittens");
imageDF.select("image.origin", "image.width", "image.height").show(false);
/*
Will output:
+-----------------------------------------------------------------------+-----+------+
|origin                                                                 |width|height|
+-----------------------------------------------------------------------+-----+------+
|file:///spark/data/mllib/images/origin/kittens/54893.jpg               |300  |311   |
|file:///spark/data/mllib/images/origin/kittens/DP802813.jpg            |199  |313   |
|file:///spark/data/mllib/images/origin/kittens/29.5.a_b_EGDP022204.jpg |300  |200   |
|file:///spark/data/mllib/images/origin/kittens/DP153539.jpg            |300  |296   |
+-----------------------------------------------------------------------+-----+------+
*/

In PySpark we provide Spark SQL data source API for loading image data as a DataFrame.

>>> df = spark.read.format("image").option("dropInvalid", True).load("data/mllib/images/origin/kittens")
>>> df.select("image.origin", "image.width", "image.height").show(truncate=False)
+-----------------------------------------------------------------------+-----+------+
|origin                                                                 |width|height|
+-----------------------------------------------------------------------+-----+------+
|file:///spark/data/mllib/images/origin/kittens/54893.jpg               |300  |311   |
|file:///spark/data/mllib/images/origin/kittens/DP802813.jpg            |199  |313   |
|file:///spark/data/mllib/images/origin/kittens/29.5.a_b_EGDP022204.jpg |300  |200   |
|file:///spark/data/mllib/images/origin/kittens/DP153539.jpg            |300  |296   |
+-----------------------------------------------------------------------+-----+------+

In SparkR we provide Spark SQL data source API for loading image data as a DataFrame.

> df = read.df("data/mllib/images/origin/kittens", "image")
> head(select(df, df$image.origin, df$image.width, df$image.height))

1               file:///spark/data/mllib/images/origin/kittens/54893.jpg
2            file:///spark/data/mllib/images/origin/kittens/DP802813.jpg
3 file:///spark/data/mllib/images/origin/kittens/29.5.a_b_EGDP022204.jpg
4            file:///spark/data/mllib/images/origin/kittens/DP153539.jpg
  width height
1   300    311
2   199    313
3   300    200
4   300    296

LIBSVM data source

This LIBSVM data source is used to load ‘libsvm’ type files from a directory. The loaded DataFrame has two columns: label containing labels stored as doubles and features containing feature vectors stored as Vectors. The schemas of the columns are:

LibSVMDataSource implements a Spark SQL data source API for loading LIBSVM data as a DataFrame.

scala> val df = spark.read.format("libsvm").option("numFeatures", "780").load("data/mllib/sample_libsvm_data.txt")
df: org.apache.spark.sql.DataFrame = [label: double, features: vector]

scala> df.show(10)
+-----+--------------------+
|label|            features|
+-----+--------------------+
|  0.0|(780,[127,128,129...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[124,125,126...|
|  1.0|(780,[152,153,154...|
|  1.0|(780,[151,152,153...|
|  0.0|(780,[129,130,131...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[99,100,101,...|
|  0.0|(780,[154,155,156...|
|  0.0|(780,[127,128,129...|
+-----+--------------------+
only showing top 10 rows

LibSVMDataSource implements Spark SQL data source API for loading LIBSVM data as a DataFrame.

Dataset<Row> df = spark.read.format("libsvm").option("numFeatures", "780").load("data/mllib/sample_libsvm_data.txt");
df.show(10);
/*
Will output:
+-----+--------------------+
|label|            features|
+-----+--------------------+
|  0.0|(780,[127,128,129...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[124,125,126...|
|  1.0|(780,[152,153,154...|
|  1.0|(780,[151,152,153...|
|  0.0|(780,[129,130,131...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[99,100,101,...|
|  0.0|(780,[154,155,156...|
|  0.0|(780,[127,128,129...|
+-----+--------------------+
only showing top 10 rows
*/

In PySpark we provide Spark SQL data source API for loading LIBSVM data as a DataFrame.

>>> df = spark.read.format("libsvm").option("numFeatures", "780").load("data/mllib/sample_libsvm_data.txt")
>>> df.show(10)
+-----+--------------------+
|label|            features|
+-----+--------------------+
|  0.0|(780,[127,128,129...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[124,125,126...|
|  1.0|(780,[152,153,154...|
|  1.0|(780,[151,152,153...|
|  0.0|(780,[129,130,131...|
|  1.0|(780,[158,159,160...|
|  1.0|(780,[99,100,101,...|
|  0.0|(780,[154,155,156...|
|  0.0|(780,[127,128,129...|
+-----+--------------------+
only showing top 10 rows

In SparkR we provide Spark SQL data source API for loading LIBSVM data as a DataFrame.

> df = read.df("data/mllib/sample_libsvm_data.txt", "libsvm")
> head(select(df, df$label, df$features), 10)

   label                      features
1      0 <environment: 0x7fe6d35366e8>
2      1 <environment: 0x7fe6d353bf78>
3      1 <environment: 0x7fe6d3541840>
4      1 <environment: 0x7fe6d3545108>
5      1 <environment: 0x7fe6d354c8e0>
6      0 <environment: 0x7fe6d35501a8>
7      1 <environment: 0x7fe6d3555a70>
8      1 <environment: 0x7fe6d3559338>
9      0 <environment: 0x7fe6d355cc00>
10     0 <environment: 0x7fe6d35643d8>