Module:TableTools

Revision as of 23:46, 19 December 2013 by imported>Mr. Stradivarius (only allow two or more arguments for the set functions)
Jump to navigation Jump to search

This module includes a number of functions for dealing with Lua tables. It is a meta-module, meant to be called from other Lua modules, and should not be called directly from #invoke.

Loading the module

To use any of the functions, first you must load the module.

<syntaxhighlight lang="lua"> local TableTools = require('Module:TableTools') </syntaxhighlight>

isPositiveInteger

<syntaxhighlight lang="lua"> TableTools.isPositiveInteger(value) </syntaxhighlight>

Returns true if value is a positive integer, and false if not. Although it doesn't operate on tables, it is included here as it is useful for determining whether a given table key is in the array part or the hash part of a table.

isNan

<syntaxhighlight lang="lua"> TableTools.isNan(value) </syntaxhighlight>

Returns true if value is a NaN value, and false if not. Although it doesn't operate on tables, it is included here as it is useful for determining whether a value can be a valid table key. (Lua will generate an error if a NaN value is used as a table key.)

shallowClone

<syntaxhighlight lang="lua"> TableTools.shallowClone(t) </syntaxhighlight>

Returns a clone of a table. The value returned is a new table, but all subtables and functions are shared. Metamethods are respected, but the returned table will have no metatable of its own. If you want to make a new table with no shared subtables and with metatables transferred, you can use mw.clone instead. If you want to make a new table with no shared subtables and without metatables transferred, use deepCopy with the noMetatable option.

removeDuplicates

<syntaxhighlight lang="lua"> TableTools.removeDuplicates(t) </syntaxhighlight>

Removes duplicate values from an array. This function is only designed to work with standard arrays: keys that are not positive integers are ignored, as are all values after the first nil value. (For arrays containing nil values, you can use compressSparseArray first.) The function tries to preserve the order of the array: the earliest non-unique value is kept, and all subsequent duplicate values are removed. For example, for the table <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{5, 4, 4, 3, 4, 2, 2, 1}</syntaxhighlight> removeDuplicates will return <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{5, 4, 3, 2, 1}</syntaxhighlight>.

numKeys

<syntaxhighlight lang="lua"> TableTools.numKeys(t) </syntaxhighlight>

Takes a table t and returns an array containing the numbers of any positive integer keys that have non-nil values, sorted in numerical order. For example, for the table <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{'foo', nil, 'bar', 'baz', a = 'b'}</syntaxhighlight>, numKeys will return <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{1, 3, 4}</syntaxhighlight>.

affixNums

<syntaxhighlight lang="lua"> TableTools.affixNums(t, prefix, suffix) </syntaxhighlight>

Takes a table t and returns an array containing the numbers of keys with the optional prefix prefix and the optional suffix suffix. For example, for the table <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{a1 = 'foo', a3 = 'bar', a6 = 'baz'}</syntaxhighlight> and the prefix 'a', affixNums will return <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{1, 3, 6}</syntaxhighlight>. All characters in prefix and suffix are interpreted literally.

numData

<syntaxhighlight lang="lua"> TableTools.numData(t, compress) </syntaxhighlight>

Given a table with keys like "foo1", "bar1", "foo2", and "baz2", returns a table of subtables in the format <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{ [1] = {foo = 'text', bar = 'text'}, [2] = {foo = 'text', baz = 'text'} }</syntaxhighlight>. Keys that don't end with an integer are stored in a subtable named "other". The compress option compresses the table so that it can be iterated over with ipairs.

compressSparseArray

<syntaxhighlight lang="lua"> TableTools.compressSparseArray(t) </syntaxhighlight>

Takes an array t with one or more nil values, and removes the nil values while preserving the order, so that the array can be safely traversed with ipairs. Any keys that are not positive integers are removed. For example, for the table <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{1, nil, foo = 'bar', 3, 2}</syntaxhighlight>, compressSparseArray will return <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{1, 3, 2}</syntaxhighlight>.

sparseIpairs

<syntaxhighlight lang="lua"> TableTools.sparseIpairs(t) </syntaxhighlight>

This is an iterator function for traversing a sparse array t. It is similar to ipairs, but will continue to iterate until the highest numerical key, whereas ipairs may stop after the first nil value. Any keys that are not positive integers are ignored.

Usually sparseIpairs is used in a generic for loop.

<syntaxhighlight lang="lua"> for i, v in TableTools.sparseIpairs(t) do

  -- code block

end </syntaxhighlight>

Note that sparseIpairs uses the pairs function in its implementation. Although some table keys appear to be ignored, all table keys are accessed when it is run.

size

<syntaxhighlight lang="lua"> TableTools.size(t) </syntaxhighlight>

Finds the size of a key/value pair table. For example, for the table <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{foo = 'foo', bar = 'bar'}</syntaxhighlight>, size will return 2. The function will also work on arrays, but for arrays it is more efficient to use the # operator. Note that to find the table size, this function uses the pairs function to iterate through all of the table keys.

keysToList

<syntaxhighlight lang="lua"> TableTools.keysToList(t, keySort, checked) </syntaxhighlight>

Returns a list of the keys in a table, sorted using either a default comparison function or a custom keySort function, which follows the same rules as the comp function supplied to table.sort. If keySort is false, no sorting is done. Set checked to true to skip the internal type checking.

sortedPairs

<syntaxhighlight lang="lua"> TableTools.sortedPairs(t, keySort) </syntaxhighlight>

Iterates through a table, with the keys sorted using the keysToList function. If there are only numerical keys, sparseIpairs is probably more efficient.

isArray

<syntaxhighlight lang="lua"> TableTools.isArray(value) </syntaxhighlight>

Returns true if value is a table and all keys are consecutive integers starting at 1.

isArrayLike

<syntaxhighlight lang="lua"> TableTools.isArrayLike(value) </syntaxhighlight>

Returns true if value is iterable and all keys are consecutive integers starting at 1.

invert

<syntaxhighlight lang="lua"> TableTools.invert(arr) </syntaxhighlight>

Transposes the keys and values in an array. For example, <syntaxhighlight lang="lua" class="" id="" style="" inline="1">invert{ "a", "b", "c" }</syntaxhighlight> yields <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{ a=1, b=2, c=3 }</syntaxhighlight>.

listToSet

<syntaxhighlight lang="lua"> TableTools.listToSet(arr) </syntaxhighlight>

Creates a set from the array part of the table arr. Indexing the set by any of the values of the array returns true. For example, <syntaxhighlight lang="lua" class="" id="" style="" inline="1">listToSet{ "a", "b", "c" }</syntaxhighlight> yields <syntaxhighlight lang="lua" class="" id="" style="" inline="1">{ a=true, b=true, c=true }</syntaxhighlight>. See also Module:Lua set for more advanced ways to create a set.

deepCopy

<syntaxhighlight lang="lua"> TableTools.deepCopy(orig, noMetatable, alreadySeen) </syntaxhighlight>

Creates a copy of the table orig. As with mw.clone, all values that are not functions are duplicated and the identity of tables is preserved. If noMetatable is true, then the metatable (if any) is not copied. Can copy tables loaded with mw.loadData.

Similar to mw.clone, but mw.clone cannot copy tables loaded with mw.loadData and does not allow metatables not to be copied.

sparseConcat

<syntaxhighlight lang="lua"> TableTools.sparseConcat(t, sep, i, j) </syntaxhighlight>

Concatenates all values in the table that are indexed by a positive integer, in order. For example, <syntaxhighlight lang="lua" class="" id="" style="" inline="1">sparseConcat{ "a", nil, "c", "d" }</syntaxhighlight> yields <syntaxhighlight lang="lua" class="" id="" style="" inline="1">"acd"</syntaxhighlight> and <syntaxhighlight lang="lua" class="" id="" style="" inline="1">sparseConcat{ nil, "b", "c", "d" }</syntaxhighlight> yields <syntaxhighlight lang="lua" class="" id="" style="" inline="1">"bcd"</syntaxhighlight>.

length

<syntaxhighlight lang="lua"> TableTools.length(t, prefix) </syntaxhighlight>

Finds the length of an array or of a quasi-array with keys with an optional prefix such as "data1", "data2", etc. It uses an exponential search algorithm to find the length, so as to use as few table lookups as possible.

This algorithm is useful for arrays that use metatables (e.g. frame.args) and for quasi-arrays. For normal arrays, just use the # operator, as it is implemented in C and will be quicker.

inArray

<syntaxhighlight lang="lua"> TableTools.inArray(arr, valueToFind) </syntaxhighlight>

Returns true if valueToFind is a member of the array arr, and false otherwise.


--[[
------------------------------------------------------------------------------------
--                               TableTools                                       --
--                                                                                --
-- This module includes a number of functions for dealing with Lua tables.        --
-- It is a meta-module, meant to be called from other Lua modules, and should     --
-- not be called directly from #invoke.                                           --
------------------------------------------------------------------------------------
--]]

local libraryUtil = require('libraryUtil')

local p = {}

-- Define often-used variables and functions.
local floor = math.floor
local infinity = math.huge
local checkType = libraryUtil.checkType

--[[
------------------------------------------------------------------------------------
-- isPositiveInteger
--
-- This function returns true if the given value is a positive integer, and false
-- if not. Although it doesn't operate on tables, it is included here as it is
-- useful for determining whether a given table key is in the array part or the
-- hash part of a table.
------------------------------------------------------------------------------------
--]]
function p.isPositiveInteger(v)
	if type(v) == 'number' and v >= 1 and floor(v) == v and v < infinity then
		return true
	else
		return false
	end
end

--[[
------------------------------------------------------------------------------------
-- isNan
--
-- This function returns true if the given number is a NaN value, and false
-- if not. Although it doesn't operate on tables, it is included here as it is
-- useful for determining whether a value can be a valid table key. Lua will
-- generate an error if a NaN is used as a table key.
------------------------------------------------------------------------------------
--]]
function p.isNan(v)
	if type(v) == 'number' and tostring(v) == '-nan' then
		return true
	else
		return false
	end
end

--[[
------------------------------------------------------------------------------------
-- removeDuplicates
--
-- This removes duplicate values from an array. Non-positive-integer keys are
-- ignored. The earliest value is kept, and all subsequent duplicate values are
-- removed, but otherwise the array order is unchanged.
------------------------------------------------------------------------------------
--]]
function p.removeDuplicates(t)
	checkType('removeDuplicates', 1, t, 'table')
	local isNan = p.isNan
	local ret, exists = {}, {}
	for i, v in ipairs(t) do
		if isNan(v) then
			-- NaNs can't be table keys, and they are also unique, so we don't need to check existence.
			ret[#ret + 1] = v
		else
			if not exists[v] then
				ret[#ret + 1] = v
				exists[v] = true
			end
		end	
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- union
--
-- This returns the union of the key/value pairs of n tables. If any of the tables
-- contain different values for the same table key, the table value is converted
-- to an array holding all of the different values.
------------------------------------------------------------------------------------
--]]
function p.union(...)
	local lim = select('#', ...) 
	if lim < 2 then
		error(lim .. ' argument' .. (lim == 1 and '' or 's') .. " passed to 'union' (minimum is 2)", 2)
	end
	local ret, trackArrays = {}, {}
	for i = 1, lim do
		local t = select(i, ...)
		checkType('union', i, t, 'table')
		for k, v in pairs(t) do
			local retKey = ret[k]
			if retKey == nil then
				ret[k] = v
			elseif retKey ~= v then
				if trackArrays[k] then
					local array = ret[k]
					local valExists
					for _, arrayVal in ipairs(array) do
						if arrayVal == v then
							valExists = true
							break
						end
					end
					if not valExists then
						array[#array + 1] = v
						ret[k] = array
					end
				else
					ret[k] = {ret[k], v}
					trackArrays[k] = true
				end
			end
		end
	end
	return ret
end				

--[[
------------------------------------------------------------------------------------
-- valueUnion
--
-- This returns the union of the values of n tables, as an array. For example, for
-- the tables {1, 3, 4, 5, foo = 7} and {2, bar = 3, 5, 6}, union will return
-- {1, 2, 3, 4, 5, 6, 7}.
------------------------------------------------------------------------------------
--]]
function p.valueUnion(...)
	local lim = select('#', ...) 
	if lim < 2 then
		error(lim .. ' argument' .. (lim == 1 and '' or 's') .. " passed to 'valueUnion' (minimum is 2)", 2)
	end
	local isNan = p.isNan
	local ret, exists = {}, {}
	for i = 1, lim do
		local t = select(i, ...)
		checkType('valueUnion', i, t, 'table')
		for k, v in pairs(t) do
			if isNan(v) then
				ret[#ret + 1] = v
			elseif not exists[v] then
				ret[#ret + 1] = v
				exists[v] = true
			end
		end
	end
	return ret
end	

--[[
------------------------------------------------------------------------------------
-- intersection
--
-- This returns the intersection of the key/value pairs of n tables. Both the key
-- and the value must match to be included in the resulting table.
------------------------------------------------------------------------------------
--]]
function p.intersection(...)
	local lim = select('#', ...) 
	if lim < 2 then
		error(lim .. ' argument' .. (lim == 1 and '' or 's') .. " passed to 'intersection' (minimum is 2)", 2)
	end
	local ret, track, pairCounts = {}, {}, {}
	for i = 1, lim do
		local t = select(i, ...)
		checkType('intersection', i, t, 'table')
		for k, v in pairs(t) do
			local trackVal = track[k]
			if trackVal == nil then
				track[k] = v
				pairCounts[k] = 1
			elseif trackVal == v then
				pairCounts[k] = pairCounts[k] + 1
			end
		end
	end
	for k, v in pairs(track) do
		if pairCounts[k] == lim then
			ret[k] = v
		end
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- valueIntersection
--
-- This returns the intersection of the values of n tables, as an array. For
-- example, for the tables {1, 3, 4, 5, foo = 7} and {2, bar = 3, 5, 6}, 
-- intersection will return {3, 5}.
------------------------------------------------------------------------------------
--]]
function p.valueIntersection(...)
	local lim = select('#', ...) 
	if lim < 2 then
		error(lim .. ' argument' .. (lim == 1 and '' or 's') .. " passed to 'valueIntersection' (minimum is 2)", 2)
	end
	local isNan = p.isNan
	local vals, ret = {}, {}
	local isSameTable = true -- Tracks table equality.
	local tableTemp -- Used to store the table from the previous loop so that we can check table equality.
	for i = 1, lim do
		local t = select(i, ...)
		checkType('valueIntersection', i, t, 'table')
		if tableTemp and t ~= tableTemp then
			isSameTable = false
		end
		tableTemp = t
		for k, v in pairs(t) do
			-- NaNs are never equal to any other value, so they can't be in the intersection.
			-- Which is lucky, as they also can't be table keys.
			if not isNan(v) then
				local valCount = vals[v] or 0
				vals[v] = valCount + 1
			end
		end
	end
	if isSameTable then
		-- If all the tables are equal, then the intersection is that table (including NaNs).
		-- All we need to do is convert it to an array and remove duplicate values.
		for k, v in pairs(tableTemp) do
			ret[#ret + 1] = v
		end
		return p.removeDuplicates(ret)
	end
	for val, count in pairs(vals) do
		if count == lim then
			ret[#ret + 1] = val
		end
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- complement
--
-- This returns the relative complement of t1, t2, ..., in tn. The complement
-- is of key/value pairs. This is equivalent to all the key/value pairs that are in
-- tn but are not in t1, t2, ... tn-1.
------------------------------------------------------------------------------------
--]]
function p.complement(...)
	local lim = select('#', ...) 
	if lim < 2 then
		error(lim .. ' argument' .. (lim == 1 and '' or 's') .. " passed to 'complement' (minimum is 2)", 2)
	end
	--[[
	-- Now we know that we have at least two sets.
	-- First, get all the key/value pairs in tn. We can't simply make ret equal to tn,
	-- as that will affect the value of tn for the whole module.
	--]]
	local tn = select(lim, ...)
	checkType('complement', lim, tn, 'table')
	local ret = {}
	for k, v in pairs(tn) do
		ret[k] = v
	end
	-- Remove all the key/value pairs in t1, t2, ..., tn-1.
	for i = 1, lim - 1 do
		local t = select(i, ...)
		checkType('complement', i, t, 'table')
		for k, v in pairs(t) do
			if ret[k] == v then
				ret[k] = nil
			end
		end
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- numKeys
--
-- This takes a table and returns an array containing the numbers of any numerical
-- keys that have non-nil values, sorted in numerical order.
------------------------------------------------------------------------------------
--]]
function p.numKeys(t)
	checkType('numKeys', 1, t, 'table')
	local isPositiveInteger = p.isPositiveInteger
	local nums = {}
	for k, v in pairs(t) do
		if isPositiveInteger(k) then
			nums[#nums + 1] = k
		end
	end
	table.sort(nums)
	return nums
end

--[[
------------------------------------------------------------------------------------
-- affixNums
--
-- This takes a table and returns an array containing the numbers of keys with the
-- specified prefix and suffix. For example, for the table
-- {a1 = 'foo', a3 = 'bar', a6 = 'baz'} and the prefix "a", affixNums will
-- return {1, 3, 6}.
------------------------------------------------------------------------------------
--]]
function p.affixNums(t, prefix, suffix)
	checkType('affixNums', 1, t, 'table')
	checkType('affixNums', 2, prefix, 'string', true)
	checkType('affixNums', 3, suffix, 'string', true)
	prefix = prefix or ''
	suffix = suffix or ''
	local pattern = '^' .. prefix .. '([1-9]%d*)' .. suffix .. '$'
	local nums = {}
	for k, v in pairs(t) do
		if type(k) == 'string' then			
			local num = mw.ustring.match(k, pattern)
			if num then
				nums[#nums + 1] = tonumber(num)
			end
		end
	end
	table.sort(nums)
	return nums
end

--[[
------------------------------------------------------------------------------------
-- compressSparseArray
--
-- This takes an array with one or more nil values, and removes the nil values
-- while preserving the order, so that the array can be safely traversed with
-- ipairs.
------------------------------------------------------------------------------------
--]]
function p.compressSparseArray(t)
	checkType('compressSparseArray', 1, t, 'table')
	local ret = {}
	local nums = p.numKeys(t)
	for _, num in ipairs(nums) do
		ret[#ret + 1] = t[num]
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- sparseIpairs
--
-- This is an iterator for sparse arrays. It can be used like ipairs, but can
-- handle nil values.
------------------------------------------------------------------------------------
--]]
function p.sparseIpairs(t)
	checkType('sparseIpairs', 1, t, 'table')
	local nums = p.numKeys(t)
	local i = 0
	local lim = #nums
	return function ()
		i = i + 1
		if i <= lim then
			local key = nums[i]
			return key, t[key]
		end
	end
end

--[[
------------------------------------------------------------------------------------
-- size
--
-- This returns the size of a key/value pair table. It will also work on arrays,
-- but for arrays it is more efficient to use the # operator.
------------------------------------------------------------------------------------
--]]
function p.size(t)
	checkType('size', 1, t, 'table')
	local i = 0
	for k in pairs(t) do
		i = i + 1
	end
	return i
end

return p