I want to have a read only table in my Lua program. If ever a key is removed or a key is associated with a new value, an error must be thrown.
function readonly(table)
local meta = { } -- metatable for proxy
local proxy = { } -- this table is always empty
meta.__index = table -- refer to table for lookups
meta.__newindex = function(t, key, value)
error("You cannot make any changes to this table!")
end
setmetatable(proxy, meta)
return proxy -- user will use proxy instead
end
It works great.
t = { }
t["Apple"] = "Red"
t[true] = "True!"
t[51] = 29
for k,v in pairs(t) do
print(v)
end
t = readonly(t)
t[51] = 30
Prints
Red
True!
29
input:7: You cannot make any changes to this table!
Problem
for k, v in pairs(t) do
print(v)
end
Will print nothing under all circumstances now. That's because the proxy table will never have anything inside of it. pairs apparently never calls index and thus cannot retrieve anything from the actual table.
What can I do to make this readonly table iterable?
I'm on Lua 5.1 and have access to these metamethods:
Lua 5.1 Manual
You can modify standard Lua function pairs to work correctly with your read-only tables.
local function readonly_newindex(t, key, value)
error("You cannot make any changes to this table!")
end
function readonly(tbl)
return
setmetatable({}, {
__index = tbl,
__newindex = readonly_newindex
})
end
local original_pairs = pairs
function pairs(tbl)
if next(tbl) == nil then
local mt = getmetatable(tbl)
if mt and mt.__newindex == readonly_newindex then
tbl = mt.__index
end
end
return original_pairs(tbl)
end
Usage:
t = { }
t["Apple"] = "Red"
t[true] = "True!"
t[51] = 29
for k,v in pairs(t) do
print(k, v)
end
t = readonly(t)
for k,v in pairs(t) do
print(k, v)
end
t[51] = 30
One solution is to create a wholly custom iterator for the table.
function readonly(table)
local meta = { } -- metatable for proxy
local proxy = { } -- this table is always empty
meta.__index = table -- refer to table for lookups
meta.__newindex = function(t, key, value)
error("You cannot make any changes to this table!")
end
local function iter()
return next, table
end
setmetatable(proxy, meta)
return proxy, iter -- user will use proxy instead
end
Usage:
t = { }
t["Apple"] = "Red"
t[true] = "True!"
t[51] = 29
for k,v in pairs(t) do
print(v)
end
t, tIter = readonly(t)
t[51] = 30
for k, v in tIter do
print(v)
end
I am trying to convert existing python function into lua function. But my lua function is not producing same result as python function. Any help is appreciated.
Python function:
import json
test = '{"http://localhost:8080/":{"phone":{"-detail/phone detail.template.html":"5167n,a,7,2","s/motorola-xoom-with-wifi.json":"516a0,5,4,3"},"favicon.ico":"016ad,3,3,2","img/phones/motorola-xoom-with-wi-fi.":{"1.jpg":"*02s,2s,4v,h3|116da,o,l,6","2.jpg":"*02s,2s,4v,kp|116da,j,i,8","3.jpg":"*02s,2s,4v,ob|116da,o,m,8,7,,7,7,7","4.jpg":"*02s,2s,4v,rx|116da,o,m,9,8,,7,7,7","5.jpg":"*02s,2s,4v,vj|116da,p,m,a,8,,7,7,7"}}}'
def tri(param):
t = {}
for key in param:
if key not in param:
continue
if isinstance(param[key], dict) and param[key] is not None:
flat = tri(param[key])
for x in flat:
if x not in flat:
continue
t[key + x] = flat[x]
else:
t[key] = param[key]
return t
print(tri(json.loads(test)))
Lua code ( which is not producing same result as python function)
local json = require('cjson')
local test = '{"http://localhost:8080/":{"phone":{"-detail/phone-detail.template.html":"5167n,a,7,2","s/motorola-xoom-with-wi-fi.json":"516a0,5,4,3"},"favicon.ico":"016ad,3,3,2","img/phones/motorola-xoom-with-wi-fi.":{"1.jpg":"*02s,2s,4v,h3|116da,o,l,6","2.jpg":"*02s,2s,4v,kp|116da,j,i,8","3.jpg":"*02s,2s,4v,ob|116da,o,m,8,7,,7,7,7","4.jpg":"*02s,2s,4v,rx|116da,o,m,9,8,,7,7,7","5.jpg":"*02s,2s,4v,vj|116da,p,m,a,8,,7,7,7"}}}'
local function tri(param)
t = {}
for key in pairs(param) do
if param[key] == nil then end
if type(param[key]) == "table" then
flat = tri(param[key])
for k in pairs(flat) do
t[key .. k] = flat[k]
end
else
t[key] = param[key]
end
end
return t
end
print(json.encode(tri(json.decode(test))))
local function tri(param)
t = {} -- every time we call tri t will be "reset" to an empty table
for key in pairs(param) do
if param[key] == nil then end
if type(param[key]) == "table" then
flat = tri(param[key]) -- here we call tri, but we still need t!
for k in pairs(flat) do
t[key .. k] = flat[k]
end
else
t[key] = param[key]
end
end
return t
end
Making at least t global should solve that problem. But there is also no reason for flat to be global so we make it local too.
local function tri(param)
local t = {}
for key in pairs(param) do
if param[key] == nil then end
if type(param[key]) == "table" then
local flat = tri(param[key])
for k in pairs(flat) do
t[key .. k] = flat[k]
end
else
t[key] = param[key]
end
end
return t
end
Your task could be done a bit easier using json.traverse() function from this Lua JSON module.
Traversing lets you perform arbitrary operations with JSON elements on-the-fly.
This code concatenates element's path (for every JSON element except JSON containers: arrays/objects) and uses it as a key for Lua table.
local json = require'json'
local t = {}
local function callback(path, json_type, value)
if value ~= nil then -- value == nil for containers (arrays/objects)
t[table.concat(path)] = value
end
end
local test = '{"http://localhost:8080/":{"phone":{"-detail/phone detail.template.html":"5167n,a,7,2","s/motorola-xoom-with-wifi.json":"516a0,5,4,3"},"favicon.ico":"016ad,3,3,2","img/phones/motorola-xoom-with-wi-fi.":{"1.jpg":"*02s,2s,4v,h3|116da,o,l,6","2.jpg":"*02s,2s,4v,kp|116da,j,i,8","3.jpg":"*02s,2s,4v,ob|116da,o,m,8,7,,7,7,7","4.jpg":"*02s,2s,4v,rx|116da,o,m,9,8,,7,7,7","5.jpg":"*02s,2s,4v,vj|116da,p,m,a,8,,7,7,7"}}}'
json.traverse(test, callback)
-- Now t == {
-- ["http://localhost:8080/favicon.ico"] = "016ad,3,3,2",
-- ["http://localhost:8080/img/phones/motorola-xoom-with-wi-fi.1.jpg"] = "*02s,2s,4v,h3|116da,o,l,6",
-- ["http://localhost:8080/img/phones/motorola-xoom-with-wi-fi.2.jpg"] = "*02s,2s,4v,kp|116da,j,i,8",
-- ["http://localhost:8080/img/phones/motorola-xoom-with-wi-fi.3.jpg"] = "*02s,2s,4v,ob|116da,o,m,8,7,,7,7,7",
-- ["http://localhost:8080/img/phones/motorola-xoom-with-wi-fi.4.jpg"] = "*02s,2s,4v,rx|116da,o,m,9,8,,7,7,7",
-- ["http://localhost:8080/img/phones/motorola-xoom-with-wi-fi.5.jpg"] = "*02s,2s,4v,vj|116da,p,m,a,8,,7,7,7",
-- ["http://localhost:8080/phone-detail/phone detail.template.html"] = "5167n,a,7,2",
-- ["http://localhost:8080/phones/motorola-xoom-with-wifi.json"] = "516a0,5,4,3"
-- }
I'm trying to implement a function that returns the first non-blank string from the variables passed to it. Unfortunately, some of these variables might be nil, so the naive approach
function first_non_empty(...)
for i, item in ipairs({...}) do
if item ~= nil and item ~= '' then
return item
end
end
return ''
end
doesn't work: ipairs quits out as soon as it encounters a nil value. This can be fixed by changing the requirements so that the variables can't be nil, or by passing the length to the function so table length doesn't have to rely on ipairs, or by wrapping all parameters in a function so that none of them are explicitly nil
function first_non_empty_func(...)
for i, func in ipairs({...}) do
local item = func()
if item ~= nil and item ~= '' then
return item
end
end
return ''
end
function fn(p)
local f = function() return p end
return f
end
-- change callers to first_non_empty_func(fn(a), fn(b), fn(c))
However, both of these solutions complicate the function prototype. Does there exist a function taking an ordered list of parameters, some of which may be nil, which returns the first of those parameters which is both non-nil and not an empty string?
Use table.pack, which preserves all nil entries and returns the number of entries in the n field:
function first_non_empty_pack(...)
local t = table.pack(...)
for i = 1, t.n do
local item = t[i]
if item ~= nil and item ~= '' then
return item
end
end
return ''
end
select('#', ...) can be used to get the number of provided arguments, so here is an alternative that doesn't use table.pack:
function first_non_empty_pack(...)
for i = 1, select('#', ...) do
local item = select(i, ...)
if item ~= nil and item ~= '' then
return item
end
end
return ''
end
Simpler approach is to use recursion. No extra tables created, etc:
function first_non_empty(item, ...)
if item ~= nil and item ~= '' then return item end
return first_non_empty(...)
end
But the list has to end with some ending marker. For example, boolean 'false', indicating there's no non-nil, nonempty strings.
I need to check if a member exists in a table that isn't at the next level, but along a path of members.
foo = {}
if foo.bar.joe then
print(foo.bar.joe)
end
this will cast an attempt to index field 'bar' (a nil value) because bar isn't defined.
My usual solution is to test the chain, piece-by-piece.
foo = {}
if foo.bar and foo.bar.joe then
print(foo.bar.joe)
end
but this can be very tedious when there are many nested tables. Are there a better way to do this test than piece-by-piece?
I don't understand what you try to mean by "along a path of members". From the example, I assume you are trying to find a value in a "subtable"?
local function search(master, target) --target is a string
for k,v in next, master do
if type(v)=="table" and v[target] then return true end
end
end
A simple example. If you use such a function, you can pass the foo table and the joe string to see if foo.*.joe exists. Hope this helps.
debug.setmetatable(nil, {__index = {}})
foo = {}
print(foo.bar.baz.quux)
print(({}).prd.krt.skrz.drn.zprv.zhlt.hrst.zrn) -- sorry ))
To search for an element that is at any level of a table, I would use a method such as this one:
function exists(tab, element)
local v
for _, v in pairs(tab) do
if v == element then
return true
elseif type(v) == "table" then
return exists(v, element)
end
end
return false
end
testTable = {{"Carrot", {"Mushroom", "Lettuce"}, "Mayonnaise"}, "Cinnamon"}
print(exists(testTable, "Mushroom")) -- true
print(exists(testTable, "Apple")) -- false
print(exists(testTable, "Cinnamon")) -- true
I think you're looking for something along these lines:
local function get(Obj, Field, ...)
if Obj == nil or Field == nil then
return Obj
else
return get(Obj[Field], ...)
end
end
local foo = {x = {y = 7}}
assert(get() == nil)
assert(get(foo) == foo)
assert(get(foo, "x") == foo.x)
assert(get(foo, "x", "y") == 7)
assert(get(foo, "x", "z") == nil)
assert(get(foo, "bar", "joe") == nil)
assert(get(foo, "x", "y") or 41 == 7)
assert(get(foo, "bar", "joe") or 41 == 41)
local Path = {foo, "x", "y"}
assert(get(table.unpack(Path)) == 7)
get simply traverses the given path until a nil is encountered. Seems to do the job. Feel free to think up a better name than "get" though.
As usual, exercise care when combining with or.
I'm impressed by Egor's clever answer, but in general I think we ought to not rely on such hacks.
See also
The 'Safe Table Navigation' patch for Lua 5.2 : http://lua-users.org/wiki/LuaPowerPatches
Lengthy discussion on this matter : http://lua-users.org/lists/lua-l/2010-08/threads.html#00519
Related technique : http://lua-users.org/wiki/AutomagicTables
I suspect something relevant has been implemented in MetaLua, but I can't find at the moment.
If I understood your problem correctly, here's one possibility:
function isField(s)
local t
for key in s:gmatch('[^.]+') do
if t == nil then
if _ENV[ key ] == nil then return false end
t = _ENV[ key ]
else
if t[ key ] == nil then return false end
t = t[ key ]
end
--print(key) --for DEBUGGING
end
return true
end
-- To test
t = {}
t.a = {}
t.a.b = {}
t.a.b.c = 'Found me'
if isField('t.a.b.c') then print(t.a.b.c) else print 'NOT FOUND' end
if isField('t.a.b.c.d') then print(t.a.b.c.d) else print 'NOT FOUND' end
UPDATE: As per cauterite's suggestion, here's a version that also works with locals but has to take two arguments :(
function isField(t,s)
if t == nil then return false end
local t = t
for key in s:gmatch('[^.]+') do
if t[ key ] == nil then return false end
t = t[ key ]
end
return true
end
-- To test
local
t = {}
t.a = {}
t.a.b = {}
t.a.b.c = 'Found me'
if isField(t,'a.b.c') then print(t.a.b.c) else print 'NOT FOUND' end
if isField(t,'a.b.c.d') then print(t.a.b.c.d) else print 'NOT FOUND' end
foo = {}
foo.boo = {}
foo.boo.jeo = {}
foo.boo.joe is foo['boo']['joe'] and so
i make next function
function exist(t)
local words = {}
local command
for i,v in string.gmatch(t, '%w+') do words[#words+1] = i end
command = string.format('a = %s', words[1])
loadstring(command)()
if a == nil then return false end
for count=2, #words do
a = a[words[count]]
if a == nil then return false end
end
a = nil
return true
end
foo = {}
foo.boo = {}
foo.boo.joe = {}
print(exist('foo.boo.joe.b.a'))
using loadstring to make temp variable. my lua ver is 5.1
remove loadstring at 5.2 5.3, instead using load
I have the following string data that I receive as input:
"route1,1234,1,no~,,route2,1234,1,no~,"
It represents two "records" of data... where each record has 4 fields.
I've built code to parse this string into it's individual columns / fields.
But the part that isn't working is when I test to see if I have any duplicates in field 2. Field 2 is the one that currently has "1234" as the value.
Here's the code:
function string:split(delimiter)
local result = { }
local from = 1
local delim_from, delim_to = string.find( self, delimiter, from )
while delim_from do
table.insert( result, string.sub( self, from , delim_from-1 ) )
from = delim_to + 1
delim_from, delim_to = string.find( self, delimiter, from )
end
table.insert( result, string.sub( self, from ) )
return result
end
local check_for_duplicate_entries = function(route_data)
local route
local route_detail = {}
local result =true
local errtxt
local duplicate = false
print("received :" ..route_data)
route = string.gsub(route_data, "~,,", "~")
route = route:sub(1,string.len(route)-2)
print("route :" ..route)
-- break up in to an array
route = string.split(route,"~")
for key, value in pairs(route) do
route_detail[key] = string.split(value,",")
end
local list_of_second_column_only = {}
for key,value in pairs(route_detail) do
local temp = value[2]
print(temp .. " - is the value I'm checking for")
if list_of_second_column_only[temp] == nil then
print("i dont think it exists")
list_of_second_column_only[key] = value[2]
print(list_of_second_column_only[key])
else
--found a duplicate.
return true
end
end
return false
end
print(check_for_duplicate_entries("route1,1234,1,no~,,route2,1234,1,no~,"))
I think where I'm going wrong is the test:
if list_of_second_column_only[temp] == nil then
I think I'm checking for key with the value temp instead of a value with the value that temp contains. But I don't know how to fix the syntax.
Also, I'm wondering if there's a more efficient way to do this.
The number of "records" i receive as input is dynamic / unknown, as is the value of the second column in each record.
Thanks.
EDIT 1
The post I was trying to use as a reference is: Search for an item in a Lua list
In the answer, they show how to test for a record in the table by value, instead of looping through the entire table...
if items["orange"] then
-- do something
end
I was playing around to try and do something similar...
This should be a bit more efficient with only one table creation and less regex matching.
The match does require that you're only interested in dups in the second field.
local function check_for_duplicate_entries(route_data)
assert(type(route_data)=="string")
local field_set = {}
for route in route_data:gmatch"([^~]*)~,?,?" do
local field = route:match",([^,]*)"
if field_set[field] then
return true
else
field_set[field] = true
end
end
return false
end
Try this. It's doing the check on the value of the second field.
I haven't looked at the efficiency.
if list_of_second_column_only[value[2]] == nil then
print("i dont think it exists")
list_of_second_column_only[value[2]] = true
print(list_of_second_column_only[value[2]])
else
--found a duplicate.
return true
end