Syntax & Structure

Learn PXL’s simple but powerful syntax for creating data transformations and text analysis expressions.

Language Overview

PXL is a focused expression language designed for:

Data filtering with complex conditions
Text analysis using AI-powered functions
Data distribution analysis with quantiles
Function chaining for multi-step transformations

Basic Expression Structure

Single Function Calls

filter(revenue > 1000)
extract("email addresses" from customer_notes)
score(review_text from "bad" to "excellent")
classify(feedback into ("positive", "negative", "neutral"))
ntile(sales_amount, 4)
word_count(description)

Function Chaining

Use the -> operator to chain function outputs:

filter(customer_type is "premium") ->
extract("product mentions" from feedback_text)

filter(order_date after "2024-01-01") ->
score(satisfaction_survey from "dissatisfied" to "satisfied")

Identifiers and Literals

Column Identifiers

Reference dataset columns by name:

customer_id
order_date
total_revenue
customer_type
product_name

Special Identifier

The $ symbol refers to the current dataset:

$  // Represents the current dataset

String Literals

Strings must be enclosed in double quotes:

"premium"
"2024-01-01"
"iPhone"
"email addresses"

Numeric Literals

Numbers support integers, decimals, and scientific notation:

42
123.45
-5
1.2e6

Temporal Literals

Dates use ISO format:

"2024-01-01"              // Date only
"2024-01-01T12:30:00"     // Date and time

Boolean Literals

true
false
TRUE    // Case insensitive
FALSE

Null Literal

null

Operators

Comparison Operators

// Numeric comparisons
revenue > 1000
age >= 18
score < 50
rating <= 5
price == 99.99
quantity != 0

// String equality
status is "active"
category is not "discontinued"

// Pattern matching
email like "%.com"
product not like "old%"

Set Operators

// Membership testing
region in ("US", "Canada", "Mexico")
status not in ("cancelled", "refunded")
priority in (1, 2, 3)

// Text containment
description contains "premium"

Temporal Operators

// Date/time comparisons
created_date after "2024-01-01"
expiry_date before "2024-12-31"
event_date between "2024-01-01" and "2024-03-31"

Boolean Operators

// Logical AND
revenue > 1000 and customer_type is "premium"

// Logical OR
status is "active" or last_login after "2024-01-01"

// Complex combinations with parentheses
age >= 18 and (country is "US" or country is "CA")

Filter Expressions

Simple Conditions

filter(active == true)
filter(revenue > 10000)
filter(customer_tier is "gold")

Complex Boolean Logic

filter(
  revenue > 5000 and
  customer_type is "enterprise" and
  (region is "US" or region is "EU")
)

filter(
  status is "active" or
  (status is "trial" and signup_date after "2024-01-01")
)

Null Handling

filter(email is not null)
filter(phone_number is null)
filter(middle_name is not "")

Text Analysis Syntax

Extract Function

extract("description" from column_name)

// Examples
extract("email addresses" from contact_info)
extract("dollar amounts" from transaction_notes)
extract("phone numbers" from customer_data)

Classify Function

// Simple categories
classify(text_column into ("category1", "category2", "category3"))

// Categories with descriptions
classify(feedback_text into (
  "positive": "happy, satisfied, great experience",
  "negative": "frustrated, disappointed, problems",
  "neutral": "factual, no strong emotion"
))

Score Function

score(text_column from "negative_pole" to "positive_pole")

// Examples
score(reviews from "terrible" to "amazing")
score(feedback from "angry" to "delighted")

Word Count Function

word_count(text_column)

// Examples
word_count(description)
word_count(customer_comments)

Data Distribution

Ntile Function

ntile(numeric_column, number_of_tiles)

// Examples
ntile(revenue, 4)      // Quartiles
ntile(age, 10)         // Deciles
ntile(score, 100)      // Percentiles

Grouping and Precedence

Parentheses for Grouping

// Boolean logic grouping
filter((age >= 18 and age <= 65) or status is "exempt")

// Function parameters
filter(region in ("US", "CA", "MX"))

// List definitions
classify(text into ("positive", "negative", "neutral"))

Operator Precedence

Parentheses () - Highest precedence
Comparison operators >, <, >=, <=, ==, !=, is, like
Set operators in, contains
Boolean AND and
Boolean OR or - Lowest precedence

// These are equivalent:
filter(age > 18 and status is "active" or type is "vip")
filter(((age > 18) and (status is "active")) or (type is "vip"))

Comments

PXL supports single-line comments:

// Filter for active premium customers
filter(status is "active" and tier is "premium")

extract("product names" from reviews)  // Extract mentioned products

Expression Chaining

Simple Chaining

data -> filter(active == true) -> ntile(revenue, 4)

Multi-Step Chaining

customer_data ->
filter(signup_date after "2024-01-01") ->
classify(feedback into ("satisfied", "neutral", "unsatisfied")) ->
filter(classify_result is "satisfied")

Syntax Rules

Case Sensitivity

Keywords are case-insensitive: AND, and, And are equivalent
Column names are case-sensitive: CustomerID ≠ customerid
String literals are case-sensitive: "Premium" ≠ "premium"

Whitespace

Whitespace is generally ignored:

// These are equivalent
filter(revenue>1000and status is"active")
filter(revenue > 1000 and status is "active")

Required Elements

Function names are required and cannot be omitted
Parentheses are required for function calls
Double quotes are required for string literals
Operators must be properly spaced in some contexts

Common Syntax Patterns

Data Filtering Patterns

// Date range filtering
filter(order_date between "2024-01-01" and "2024-12-31")

// Multi-condition filtering
filter(amount > 100 and currency is "USD" and status is "paid")

// Category filtering
filter(product_category in ("electronics", "books", "clothing"))

Text Analysis Patterns

// Sentiment analysis workflow
filter(review_date after "2024-01-01") ->
score(review_text from "disappointed" to "thrilled")

// Content classification
filter(content_type is "support_ticket") ->
classify(message_text into ("technical", "billing", "general"))

Data Exploration Patterns

// Revenue quartile analysis
filter(customer_type is "enterprise") ->
ntile(annual_revenue, 4)

// Text length analysis
filter(has_description == true) ->
word_count(product_description)

Error Prevention

Quote Strings Properly

// ❌ Incorrect
filter(status is active)

// ✅ Correct
filter(status is "active")

Use Correct Date Format

// ❌ Incorrect
filter(date after "01/15/2024")

// ✅ Correct
filter(date after "2024-01-15")

Include Function Parentheses

// ❌ Incorrect
filter revenue > 1000

// ✅ Correct
filter(revenue > 1000)

Match Parentheses and Quotes

// ❌ Incorrect - missing closing quote
filter(status is "active)

// ❌ Incorrect - missing closing parenthesis
filter(revenue > 1000

// ✅ Correct
filter(status is "active")

What’s Next?

Function Reference

Learn the details of each PXL function with examples and best practices.

Function Reference

Advanced Techniques

Master complex patterns and optimization strategies for sophisticated analysis.

Advanced Guide