# Querying Lists (Arrays) & Objects (JSON) in Airscript

Sean Lynch

Airscript is designed to easily facilitate real-world information retrieval and manipulation. In the real world data is messy, doesn’t align with the user interface, and you may not be able to query it like a database. We’ll cover three common operations on data.

• Filtering: There are several techniques to search for data.
• Joining: Data of one type is often related to data of another. Querying data by those relations is a powerful technique.
• Transformations: External systems may represent their data differently from your application. Airscript has built-in features to help transform one representation into another.

There are two primary ways to work with collections of data in Airscript: Path Expressions and Query Expressions. While the same task can often be accomplished with either, they both have complementary strengths and weaknesses. We'll cover how to leverage both to retrieve and manipulate data. Where possible examples with be given in both expression styles.

For the following examples, let us consider that we have two variables, one containing a list_of_books and another that holds a selected_book.

list_of_books = [
{
title: "Angels & Demons",
author: "Dan Brown",
isbn: "9781416524793",
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
},
{
title: "Hackers: Heroes of the Computer Revolution",
author: "Steven Levy",
isbn: "9780141000510",
genres: ["History"]
},
{
title: "Moonwalking with Einstein: The Art and Science of Remembering Everything",
author: "Joshua Foer",
isbn: "9781594202292",
genres: ["Nonfiction"]
},
...
]
selected_book = {
title: "Angels & Demons"
author: "Dan Brown"
isbn: "9781416524793"
genres: ["Fiction"]
}

Note: Though the techniques in this article are quite useful, if filtering or joining can be done in Data Operation it should be done there. The external system will be more performant than doing it in Airscript yourself.

## Path Expression

The most straightforward and common way to work with Lists and Objects is to retrieve a single item by index or field name respectively. For example, in order to retrieve the author of the selected_book, use the following Path Expression

selected_book.author => "Dan Brown"

This syntax is valid for any expression, we could perform the same operation on an object literal.

{
title: "Angels & Demons"
author: "Dan Brown"
isbn: "9781416524793"
genres: ["Fiction"]
}.author => "Dan Brown"

While retrieving data from an object requires using a field name, retrieving data from a list uses a number (or index) indicating the position in the list. In order to retrieve the first book from list_of_books retrieve the item at index 0.

list_of_books[0] => {
title: "Angels & Demons"
author: "Dan Brown"
isbn: "9781416524793"
genres: ["Fiction"]
}

An important note is that these operations can be chained. To retrieve the author of the first book we first retrieve the first book and then retrieve the author of that book.

list_of_books[0].author => "Dan Brown"

And to retrieve the first genre of the first book:

list_of_book[0].genres[0] => "Fiction"

## Query Expression

The Query Expression is quite similar to SQL, or LINQ. It is used to select multiple values from a collection. The most basic Query Expression does nothing and simply returns the original collection.

FROM book IN list_of_books
SELECT book
=> [
{
title: "Angels & Demons",
author: "Dan Brown",
isbn: "9781416524793",
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
},
{
title: "Hackers: Heroes of the Computer Revolution",
author: "Steven Levy",
isbn: "9780141000510",
genres: ["History"]
},
{
title: "Moonwalking with Einstein: The Art and Science of Remembering Everything",
author: "Joshua Foer",
isbn: "9781594202292",
genres: ["Nonfiction"]
},
...
]

This works by creating a variable to be used for each item in the collection, here we assign each book in list_of_books to the variable book. The Query Expression will evaluate the SELECT clause for each book in list_of_books and return a new list with the result of the SELECT clause. While the last expression is simple, it is also rather useless. Query Expressions can do much more interesting things. We can retrieve a list of the ISBNs by modifying the SELECT clause to return each book’s ISBN.

FROM book IN list_of_books
SELECT book.isbn
=> [
"9781416524793",
"9781416524793",
"9781594202292",
"9780141000510",
...
]

Perhaps we would like to order the books alphabetically by author, we can do so by including an ORBER BY clause

FROM book IN list_of_books
ORDER BY book.author
SELECT book
=> [
{
title: "Angels & Demons",
author: "Dan Brown",
isbn: "9781416524793",
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
},
{
title: "Moonwalking with Einstein: The Art and Science of Remembering Everything",
author: "Joshua Foer",
isbn: "9781594202292",
genres: ["Nonfiction"]
},
{
title: "Hackers: Heroes of the Computer Revolution",
author: "Steven Levy",
isbn: "9780141000510",
genres: ["History"]
},
...
]

Using the LIMIT clause we could select just the first two books.

FROM book IN list_of_books
LIMIT 2
SELECT book
=> [
{
title: "Angels & Demons",
author: "Dan Brown",
isbn: "9781416524793",
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
}
]

Note that in the last examples the SELECT clause didn’t do much other than simply selecting each item. In these cases, the SELECT clause can be omitted.

FROM book IN list_of_books LIMIT 2

## Filtering

It is often necessary to search through a data set for a particular item or set of items. If we would like to search the list_of_books by a specific author, we could use a Query Expression with a WHERE clause requiring that the author field of the book be equal to “Dan Brown”.

FROM book IN list_of_books
WHERE book.author = "Dan Brown"
SELECT book
=> [
{
title: "Angels & Demons"
author: "Dan Brown"
isbn: "9781416524793"
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
}
]

The equivalent Path Expression returns the same result and looks like this:

list_of_books[?(@.author = "Dan Brown")]

There are often queries that are only intended to return a single result. For example, assuming list_of_books does not contain duplicates seaching for a book by ISBN will only have one result. To access this singular book directly we can filter based on books that have a matching ISBN. Since there should only be one, we can simply take the first element of the resulting list.

(
FROM book IN list_of_books
WHERE book.isbn = "9781416524793"
SELECT book
)[0]
=> {
title: "Angels & Demons",
author: "Dan Brown",
isbn: "9781416524793",
genres: ["Fiction"]
}

The equivalent Path Expression:

list_of_books[?(@.isbn = "9781416524793")][0]

Note that if no book is found, the result of these expressions will return NULL.

In addition to removing items from the collection, Airscript also makes it quite convenient to work with nested data. Notice that if we wanted to search by genre we could not do the same thing because each book can have more than one genre. Rather than using the equality operator, we can use the CONTAINS function to determine if the list of genres contains the genre we are looking for.

FROM book IN list_of_books
WHERE CONTAINS(book.genres, "Non-fiction")
SELECT book
=> [
{
title: "Angels & Demons"
author: "Dan Brown"
isbn: "9781416524793"
genres: ["Fiction"]
},
{
title: "The Da Vinci Code",
author: "Dan Brown",
isbn: "9780307879257",
genres: ["Fiction"]
}
]

And the equivalent Path Expression

list_of_books[?(CONTAINS(@.genres, "Non-fiction"))]

The concept of filtering can still be pushed further. Suppose we would like to retrieve a list of all the genres our list_of_books covers. We could begin by retrieving the list of genres from each book using the techniques we have seen so far. The Query Expression is fairly straightforward.

FROM boox IN list_of_books
SELECT book.genres

The following Path Expression will return the same result.

list_of_books[*].genres

However, this gives us a list of lists of genres

[
["Fiction"]
["Fiction"]
["History"]
["Nonfiction"]
...
]

Using the FLAT function we can convert this double list into a single list.

FLAT(list_of_books[*].genres)
=>
[
"Fiction"
"Fiction"
"History"
"Nonfiction"
...
] "

Finally, notice that we have duplicates. We can remove the duplicates using the DISTINCT keyword of a Query Expression

FROM genre
IN FLAT(FROM book IN list_of_books SELECT book.genres)
SELECT DISTINCT genre=> ["Fiction", "History", "Nonfiction", ...]

Note that there is no equivalent Path Expression to a DISTINCT Query Expression, however, we are free to mix and match either type of expression.

FROM genre IN FLAT(list_of_books[*].genres)
SELECT DISTINCT genre
=> ["Fiction", "History", "Nonfiction", ...]

## Joining

The next technique we will discuss is useful when relating one data set to another. We would like to add a Web Page to our application that shows the user the distinct list of genres we just collected with a description of each genre. For the following example assume we have a variable named list_of_genres that contains a list of genre objects that contain a name and a description.

list_of_genres = [
{
name: "Fiction",
description: "..."
},
{
name: "History",
description: "...",
},
{
name: "Nonfiction",
description: "...",
},
{
name: "Mystery",
description: "...",
}

We begin by using our previous expression to extract the distinct genre names from list_of_books. Rather than selecting the genre name itself, we instead filter list_of_genres by each distinct genre and return the genre that was found.

FROM genre_name IN FLAT(list_of_books[*].genres)
SELECT DISTINCT (  FROM genre IN list_of_genres
WHERE genre.name = genre_name
SELECT genre)[0]
=> [
{
name: "Fiction",
description: "..."
},
{
name: "History",
description: "...",
},
{
name: "Nonfiction",
description: "...",
}
]

The equivalent Path Expression

FROM genre IN FLAT(list_of_books[*].genres)
SELECT DISTINCT list_of_genres[?(@.name = genre_name)][0]
=> [
{
name: "Fiction",
description: "..."
},
{
name: "History",
description: "...",
},
{
name: "Nonfiction",
description: "...",
}
]

## Transformations

When querying or sending data to an external system the way you represent data may not be the same as the way an external system represents data. For example, imagine there is a service that will return book information given a list of ISBNs. However, this external system requires each ISBN to be prefixed with the string “ISBN:”. In order to send list_of_books we must not only extract the ISBN but also add the required prefix turning “9781416524793” into “ISBN:9781416524793”. Since SELECT clauses are just Airscript expressions we can use the CONCAT function to do this.

FROM book IN list_of_books
SELECT CONCAT("ISBN:", book.isbn)=> [   "ISBN:9781416524793",   "ISBN:9781416524793",   "ISBN:9781594202292",   "ISBN:9780141000510",   ...]

As we have seen we can retrieve a book with a specific ISBN by filtering our list to only the book with the matching ISBN, and then selecting the first element in that list.

(FROM book IN list_of_books WHERE book.isbn = "9781416524793")[0]=> {   title: "Angels & Demons",  author: "Dan Brown",  isbn: "9781416524793",   genres: ["Fiction"] }
list_of_books[?(@.isbn = "9781416524793")][0]

If our application has to do this often it could get tedious. If we were to create an object where the fields of the object were the ISBNs themselves we could use a simple Path Expression to retrieve the book by ISBN directly without having to filter.

books_by_isbn["9781416524793"]

The Query Expression has one more feature we haven’t explored yet. In addition to creating lists, it can also be used to create objects. In order to do so rather than an expression in the SELECT clause use a key: value pair just as you would when specifying the properties of an object.

FROM book IN list_of_books
SELECT "{{book.isbn}}": book

Note: If you were to create a variable of this type, you would choose The Any (JSON) Variable Data Type. You cannot create an App Object for this variable because the field names cannot be known in advance.