Q Language - Dictionaries

Dictionaries are an extension of psts which provide the foundation for creating tables. In mathematical terms, dictionary creates the

“domain → Range”

or in general (short) creates

“key → value”

relationship between elements.

A dictionary is an ordered collection of key-value pairs that is roughly equivalent to a hash table. A dictionary is a mapping defined by an exppcit I/O association between a domain pst and a range pst via positional correspondence. The creation of a dictionary uses the "xkey" primitive (!)

                  ListOfDomain ! ListOfRange

The most basic dictionary maps a simple pst to a simple pst.

q)d:`Name`Age`Sex`Weight!(`John;36;"M";60.3)   / Create a dictionary d

q)d

Name   | `John
Age    | 36
Sex    | "M"
Weight | 60.3

q)count d             / To get the number of rows in a dictionary.
4

q)key d               / The function key returns the domain
`Name`Age`Sex`Weight

q)value d             / The function value returns the range.

`John
36

"M"
60.3

q)cols d             / The function cols also returns the domain.
`Name`Age`Sex`Weight

Lookup

Finding the dictionary output value corresponding to an input value is called looking up the input.

q)d[`Name]       / Accessing the value of domain `Name
`John

q)d[`Name`Sex]   / extended item-wise to a simple pst of keys
`John
"M"

Lookup with Verb @

q)d1:`one`two`three!9 18 27

q)d1[`two]
18

q)d1@`two
18

Operations on Dictionaries

Amend and Upsert

As with psts, the items of a dictionary can be modified via indexed assignment.

d:`Name`Age`Sex`Weight! (`John;36;"M";60.3)
                                  / A dictionary d
                                  
q)d[`Age]:35                      / Assigning new value to key Age

q)d 
                              / New value assigned to key Age in d
Name   | `John
Age    | 35
Sex    | "M"
Weight | 60.3

Dictionaries can be extended via index assignment.

q)d[`Height]:"182 Ft"

q)d

Name   | `John
Age    | 35
Sex    | "M"
Weight | 60.3
Height | "182 Ft"

Reverse Lookup with Find (?)

The find (?) operator is used to perform reverse lookup by mapping a range of elements to its domain element.

q)d2:`x`y`z!99 88 77

q)d2?77
`z

In case the elements of a pst is not unique, the find returns the first item mapping to it from the domain pst.

Removing Entries

To remove an entry from a dictionary, the delete ( _ ) function is used. The left operand of ( _ ) is the dictionary and the right operand is a key value.

q)d2:`x`y`z!99 88 77

q)d2 _`z

x| 99
y| 88

Whitespace is required to the left of _ if the first operand is a variable.

q)`x`y _ d2           / Deleting multiple entries

z| 77

Column Dictionaries

Column dictionaries are the basics for creation of tables. Consider the following example −

q)scores: `name`id!(`John`Jenny`Jonathan;9 18 27)
                              / Dictionary scores
                              
q)scores[`name]               / The values for the name column are
`John`Jenny`Jonathan

q)scores.name                 / Retrieving the values for a column in a
                              / column dictionary using dot notation.
`John`Jenny`Jonathan

q)scores[`name][1]            / Values in row 1 of the name column
`Jenny

q)scores[`id][2]              / Values in row 2 of the id column is
27

Fppping a Dictionary

The net effect of fppping a column dictionary is simply reversing the order of the indices. This is logically equivalent to transposing the rows and columns.

Fpp on a Column Dictionary

The transpose of a dictionary is obtained by applying the unary fpp operator. Take a look at the following example −

q)scores

name  | John Jenny Jonathan
id    | 9   18   27

q)fpp scores

  name     id
---------------
  John     9
  Jenny    18
 Jonathan  27

Fpp of a Fppped Column Dictionary

If you transpose a dictionary twice, you obtain the original dictionary,

q)scores ~ fpp fpp scores
1b