Splitrows Transform

Splits a column of values into separate rows of data based on the specified delimiter. You can split rows only on String literal values. Pattern-based row splitting is not supported.

NOTE: The splitrows transform must be the first one in your recipe. When a dataset is loaded for the first time in the Transformer page, a splitrows transform may added as the first step of the recipe. You cannot cannot add another splitrows transform later in your recipe. For more information, see Initial Parsing Steps.

Basic Usage

If you load CSV data into the Transformer page and then review the first recipe step in the Recipe panel, it might look like the following:

splitrows col: column1 on: '\r'

Output: The above splits all of the CSV data, which is stored as a comma-separated values in column1 initially. The delimiter for the end of the row is a carriage return, which is indicated by the \r escaped value.

Parameters

splitrows col:column_ref on:'string_literal' [quote:'quoted_string']

TokenRequired?Transform BuilderData TypeDescription
splitrowsY

Split into rows

NOTE: This transform must be the first in the dataset. It is only available for raw datasets with all steps removed.

transformName of the transform
colYColumnstringSource column name
onYOn valuestringSpecifies the end of row delimiter for each value in the source column
quoteNIgnore matches betweenstringSpecifies a quoted object that is omitted from pattern matching
quoteEscapeCharNQuote Escape CharacterstringSpecifies the escape character that is used to precede quote marks.

For more information on syntax standards, see Language Documentation Syntax Notes.

col

Identifies the column to which to apply the transform. You can specify only one column.

Usage Notes:

Required?Data Type
YesString (column name)

on

Identifies the pattern to match. Pattern can be a string literal, Cloud Dataprep pattern, or regular expression pattern.

NOTE: Value must be a string. For this transform, the parameter defines the string on which to split the current row and add the data after the string to the new row.

Usage Notes:

Required?Data Type
YesString literal

quote

Can be used to specify a string as a single quoted object.

NOTE: This parameter value must be a single character.

splitrows col: MyCol on: '\r\n' quote: '"'

Output: Splits the MyCol column into separate rows on the return-newline character string (\r\n). Values contained within double quotes (") are treated as strings, even if they contain \r\n values.

Usage Notes:

Required?Data Type
NoString

quoteEscapeChar

By default, the platform assumes the following characters are used to escape quote marks in text-based formats that use quotes to identify fields:

  • JSON: Platform assumes that \ is used.
  • All other file formats: Platform assumes that " is used.

Optionally, you can specify the character that is used to escape quote marks in each recipe. Typically, this value is specified for processing JSON data or for customizing the transform for your specific data.

splitrows col: MyCol on: '\r\n' quote: '"' quoteEscapeChar:'"'

Usage Notes:

Required?Data Type
NoString literal (single character)

Examples

Example - splitrows with CSV data

Raw source:

Before you import, your raw data in CSV format looks like the following:

Date,UserId,Message
3/14/16,jjones,"Hi, everyone! 
Happy, St. Patrick's Day!"
3/14/16,lsmith,"@jjones, it's on 3/17."
3/14/16,thughes,lol
3/14/16,jjones,"@lsmith, no harm in celebrating twice!"

Notes:

  • The Message value for the first row of data contains carriage returns, which must be captured in the data value and not used to split the row.
  • The Message value for thughes is a single unquoted value.

Transform:

When the data is first loaded into the Transformer page, the following step is added as the first step to the recipe:

splitrows col: column1 on: '\r' quote: '\"'

This transform splits the raw CSV data on the carriage return. However, values that are stored between double quotes are treated as single strings, and no row breaks are applied to this data.

Results:

For CSV data, this step, a split step, and a header step are typically added automatically as the first steps of the recipe. In the Transformer page, this dataset should look like the following:

DateUserIdMessage
3/14/16jjonesHi, everyone! C R C RHappy, St. Patrick's Day!
3/14/16lsmith@jjones, it's on 3/17.
3/14/16thugheslol
3/14/16jjones@lsmith, no harm in celebrating twice!

The C R marker is used to indicate a carriage return in the data.


Was this page helpful? Let us know how we did:

Send feedback about...

Google Cloud Dataprep Documentation