String Functions Explained

Introduction

STRINGs are a crucial part of any dataset and being able to confidently manipulate and transform them can make all the difference in your analysis. This notebook covers the common STRING manipulations in BigQuery.

The Basics

STRINGs are their own data type in Big Query. They are the most flexible type - often our dates if not formatted exactly right will be read in as a STRING, or our survey results will be listed as STRINGs of responses. Therefore it's crucial to know how to master them.

Manipulating Strings

How to modify and transform entire STRINGs. Each of these functions returns a STRING.

Removing and adding characters

There are several functions available in BigQuery to add and remove whitespace to your STRINGs. They either involve:

• trimming: removing whitespace
• padding: adding whitespace

For trimming, the functions are:

• TRIM(value1[, value2]) -> Removes all leading and trailing characters that match value2 (whitespace if not specified)
• LTRIM(value1[, value2]) -> Removes only leading characters that match value2 (whitespace if not specified)
• RTRIM(value1[, value2]) -> Removes only trailing characters that match value2 (whitespace if not specified)

For padding, the functions available are:

• RPAD(original_value, return_length[, pattern]) -> Returns the original_value appended with the pattern up to the return_length number of characters.
• LPAD(original_value, return_length[, pattern])-> Returns the original_value prepended with the pattern up to the return_length number of characters.

Changing case

A common way to re-format STRINGs is to change the case. To do that in BigQuery, you can use:

• LOWER(value)-> Returns value in lowercase
• UPPER(value) -> Returns value in uppercase
• INITCAP(value[, delimiters]) -> Returns the first character in each word as uppercase and the rest as lowercase

Re-arranging values

To re-arrange the characters in a STRING, the most common function is:

• REVERSE(value)-> Returns string in reverse order

Analysing Strings

How to get information about each STRING. These functions may return INT64 or BOOL.

String length

To find the length of a STRING, there are a surprising number of options in BigQuery:

• BYTE_LENGTH(value) -> Returns the length of the STRING in BYTEs
• CHAR_LENGTH(value) or CHARACTER_LENGTH(value) -> Returns the length of the STRING in characters
• LENGTH(value) -> Returns number of characters

Comparing strings

When building filters or CASE statements, it's common to check whether a STRING is or is similar to another string or substring. To do that we can use one of the STRING comparison functions:

• STARTS_WITH(value1, value2)-> Returns True/False if value1 starts with the substring value2
• ENDS_WITH(value1, value2)-> Returns True/False if value1 ends with with the substring value2
• REGEXP_CONTAINS(value, regexp)-> Returns True/False if value contains the pattern matched in the regexp expression

in combination with one of the following comparison operators:

• IN, NOT IN
• LIKE, NOT LIKE

Substrings

How to find, extract, and modify substrings, or parts of STRINGs. These functions may return INT64, ARRAY or STRING.

Finding substrings

When dealing with substrings you often want to first locate a substring within a STRING. To do that, we can use:

SELECT
  CONCAT('Hello', " ", "World")

• REGEXP_INSTR(source_value, regexp [, position[, occurrence, [occurrence_position]]])-> Returns the lowest 1-based index of a regular expression, regexp, in source_value
• STRPOS(value1, value2)-> Returns the 1-based index of the first occurrence of value2 inside value1. Returns 0 if value2 is not found.

These functions tell you which index the substring first appears, useful for replacing or extracting substrings given this index.

Replacing substrings

Once you've found a substring, you likely want to either extract it or replace it. To replace a substring, you can use one of the following:

SELECT
  ARRAY_LENGTH(REGEXP_EXTRACT_ALL("how many a's in this sentence?", r'(a)')) AS a_count

• REPLACE(original_value, from_value, to_value)->Replaces all occurrences of from_value with to_value in original_value.
• REGEXP_REPLACE(value, regexp, replacement) -> Returns a STRING where all substrings of value that match regular expression regexp are replaced with replacement.

SELECT
REPLACE('My Name is ____', '____', 'Judge') AS replace,
REGEXP_REPLACE('My Name is ____', r'(_+)', 'Judge') AS regex_replace

Extracting substrings

Perhaps the most valuable manipulation to perform on substrings is to extract them to their own column or entity. To do that, there are several options:

• LEFT(value, length)-> Returns a STRING value that consists of the specified number of leftmost characters or bytes from value
• RIGHT(value, length)-> Returns a STRING value that consists of the specified number of rightmost characters or bytes from value
• SPLIT(value[, delimiter]) -> Returns an ARRAY of STRINGs split by the delimiter (or space if omitted)
• SUBSTR(value, position[, length]) -> Returns a substring of value from the position index up to the length of characters specified
• REGEXP_EXTRACT(value, regexp[, position[, occurrence]]) or REGEXP_SUBSTR(value, regexp[, position[, occurrence]]) -> Returns a substring of value that matches the regexp expression starting at position index.
• REGEXP_EXTRACT_ALL(value, regexp)-> Returns an ARRAY of substrings that match the regexp expression

Combining substrings

To combine substrings together you can use:

• CONCAT(value1[, ...])-> Concatenates 2 or more STRINGs into a single result.

How Do I...

Count the number of occurrences of a character in a string?

To do this we can make use of REGEXP_EXTRACT_ALL and ARRAY_LENGTH.

Extract numbers from string using regex?

To do this we can again use REGEXP_EXTRACT_ALL

We can choose to make our regex greedy, meaning once it found one number it will look for another one. This is the difference between:

r'([0-9]+)' : which says find 1 or more digits together and r'([0-9])' which says find any digits. Depending on what you want, both can be useful.