We Keep Coding

Using JS to sum a column of values from an external text file containing donation histories of a db of donors

I need some assistance figuring out how to sum a column of dynamic totals that could be a positive or negative dollar amount, or an indication of stock shares.
I have a tab-delimited text file of donor contributions for that I am matching up against a CSV file of other related customer data that I am using to create a statement letter which will show a "donation history" of a particular donor. Each donor has a different amount of donations, and to complicate things, the column of data for a particular donation record could show either "$1,000.00" or "($1,000.00)" or "2 Shares APPL". The number with the parentheticals is of course, representing a negative number.
At the end of this column, I need to show a string that will read either "Total: $1,000.00," or if any of the donation history contains a donation record that included shares of stock the returned string will simply read, "$1,000.00 & Stock."
I have been racking my brain trying to come up with the JS rule that can achieve this. I have the JS rule that is generating the donation history correctly, but summing the donation amount column is causing me to go crazy...
Here is the JS for generating my donation history list in the letter (this seems to be working fine):
var contributionList = new ExternalDataFileEx("/~wip/248839 Frontiers/Master Data/Double Data proof.txt", "\t");
var donor_id = Field("Supporter");
var lb = "<br>\n";
var matches = new Array();
for (var i = 0; i <= contributionList.recordCount; i++) {
var idVariable = contributionList.GetFieldValue(i, "Supporter");
var dateVariable = contributionList.GetFieldValue(i, "Donation Date");
var ministryVariable = contributionList.GetFieldValue(i, "Ministry Designation");
var giftVariable = contributionList.GetFieldValue(i, "Donation Amount");
var tsSettings = "<p tabstops=19550,Right,,;29600,Left,,;>";
var ts = "<t>";
if (donor_id == idVariable)
matches.push(tsSettings + dateVariable + ts + giftVariable + ts + ministryVariable);
}
//return matches;
return matches.join(lb);
Now here is the JS code that is not working just fine. I am trying to tally the donation amount column, it only returns "Total: $0.00 & Stock" every time (I have tried to explain my thought process via comments):
var contributionList = new ExternalDataFileEx("/~wip/248839 Frontiers/Master Data/Double Data proof.txt", "\t");
var donor_id = Field("Supporter");
for (var i = 0; i <= contributionList.recordCount; i++) {
var idVariable = contributionList.GetFieldValue(i, "Supporter");
var giftVariable = contributionList.GetFieldValue(i, "Donation Amount");
var sum = 0;
var shares = 0;
var tsSettings = "<p tabstops=19550,Right,,;29600,Left,,;>";
var ts = "<t>";
var totalStr = "Total ";
var stockStr = " & Stock";
var totalFormatted = FormatNumber("$#,###.00", Math.max(0, StringToNumber(sum)));
// Match data from linked file to current Supporter
if (donor_id == idVariable) {
// Look at current record and see if it contains the word "Share(s)"
// or not and act accordingly
if (giftVariable.match(/(^|\W)share($|\W)/i) || giftVariable.match(/(^|\W)shares($|\W)/i)) {
// Turn switch "on" if donation amount is a share or shares so
// we can have the " & Stock" appended to our string.
shares = 1;
// Because this donation is/are shares, we must "zero" this
// amount to make the math work when we sum everything up...
giftVariable = 0;
// This is where we are keeping our running total...
sum += giftVariable[i];
} else {
// This record was not a donation of share(s) so we now have to
// determine whether we are dealing with postive or negative numbers
// and then strip out all of the non-number characters, remove and
// replace the () whis just a "-," leaving us with a number we can
// work with...
// If number has parenthesis, then deal with it...
if (giftVariable.indexOf("(")) {
// Strip out all the ()$, characters...
giftVariable = giftVariable.replace(/[()$,]/g,"")
// Append the minus sign to the number...
giftVariable = "-" + giftVariable;
sum += giftVariable[i];
} else {
giftVariable = giftVariable.replace(/[$,]/g,"");
sum += giftVariable[i];
}
}
}
}
// Return Total...
if (shares == 1) {
return tsSettings + totalStr + ts + totalFormatted + stockStr;
} else {
return tsSettings + totalStr + ts + totalFormatted;
}
Any assistance would be greatly appreciated!

The problem (and code) needs to be broken into smaller, atomic steps. From your description it sounds like you should:
load a text file into memory
for each line in the file
extract: {
donor_id
charity
gift
and store the results in a contributions dictionary
for each item in the contributions dictionary
transform gift string into {
dollarAmount: float with a default of 0.0
stock: name with a default of ""
}
create an empty dictionary called totals
each item will have the shape {
id
dollarAmount as a float
stocks an an array
}
for each item in the contributions dictionary
lookup the id in the totals dictionary
if it exists
totals[id].dolarAmount += item.dollarAmount
totals[id].stocks.push(item.stock)
otherwise
totals[id].dollarAmount = item.dollarAmount
totals[id].stocks = [item.stock]
normalize your charities
for each item in totals dictionary
remove any empty strings from item.charities
create your report
for each item in totals dictionary
write`${item.id} donated `${item.dollarAmont}` ${item.stocks.length > 1 ? 'and stock' : ''
I believe you are trying to do too many things at once. Instead, the goal should be to normalize your data before you attempt to perform any calculations or aggrgrations, then normalize your aggregrations before writing your summaries or reports.
I would also stay away from using any direct string manipulation. You should have a dedicated function whose only purpose is to take a string like "($20.34) and 1 share of APPL" and return either 20.34, -20.34, or 0.0. And a different function whose only purpose is to take the same string and return either true or false is stock was present.

How to extract a specific text from a string. The hard part is the desired text changes periodically

I have an HTML document which contains this text somewhere in it
function deleteFolder() {
var mailbox = "CN=John Urban,OU=Sect-1,DC=TestServer ,DC=acme,DC=com";
var path = "/Inbox/";
//string of interest: "CN=John Urban,OU=Sect-1,DC=TestServer ,DC=acme,DC=com"
I just want to extract this text and store it in a variable in C#. My problem is that string of interest will slightly change each time the page is loaded, something like this:
"CN=John Urban,OU=Sect-1,DC=TestServer ,DC=acme,DC=com"
"CN=Jane Doe,OU=Sect-1,DC=TestServer ,DC=acme,DC=com"
etc....
How do I extract that ever changing string, without regular expression?

Is it always a function deleteFolder() which has its first line as var mailbox = "somestring"? And you are interested in somestring?
Based on the requirements you told us, could just search your string containing the HTML for var mailbox =" and then the next " and take all text between these two occurrences.
var htmlstring= "..."; //
var i1 = htmlstring.IndexOf("var mailbox = \"");
var i2 = i1 >= 0 ? htmlstring.IndexOf("\"", i1+15) : -1;
var result = i2 >= 0 ? htmlstring.Substring(i1+15, i2-(i1+15)): "not found";
VERY, VERY ugly, not maintainable, but without more information, I can't do any better. However Regex would be much nicer!

Split long string into text chunks with jQuery

I have a long string that needs to be sliced into separated chunks inside an array, with a predefined length limit the chunks. Some rules apply:
If the limit cuts a word, then the word is separated for the next chunk.
Slices must be trimmed (no spaces at the beginning or end of the array item).
Special punctuation .,!? should stay with the word, and not be sent to the next chunk.
Original text: I am totally unappreciated in my time. You can run this whole park from this room with minimal staff for up to 3 days. You think that kind of automation is easy? Or cheap? You know anybody who can network 8 connection machines and debug 2 million lines of code for what I bid for this job? Because if he can I'd like to see him try.
Result with current code ["I am totally", " unappreciated in my time", ". You can run this whole", " park from this room with", " minimal staff for up to ", "3 days. You think that", " kind of automation is ea", "sy? Or cheap? You know", " anybody who can network ", "8 connection machines", " and debug 2 million line", "s of code for what I bid", " for this job? Because if", " he can I'd like to see h", "im try."]
...it should actually be:
["I am totally", "unappreciated in my time.", "You can run this whole", "park from this room with", "minimal staff for up to 3", "days. You think that kind", "of automation is easy?", "Or cheap? You know anybody", "who can network 8", "connection machines and", "debug 2 million lines of", "code for what I bid for", "this job? Because if he", "can I'd like to see him", "try."]
As you can see, I'm still having trouble with rules 2 and 3.
This is my current code (you can check the working demo in jsfiddle):
function text_split(string, limit, pos, lines) {
//variables
if(!pos) pos = 0;
if(!lines) lines = [];
var length = string.val().length;
var length_current;
//cut string
var split = string.val().substr(pos, limit);
if(/^\S/.test(string.val().substr(pos, limit))) {
//check if it is cutting a word
split = split.replace(/\s+\S*$/, "");
}
//current string length
length_current = split.length;
//current position
pos_current = length_current + pos;
//what to do
if(pos_current < length) {
lines.push(split);
return text_split(string, limit, pos_current, lines);
} else {
console.log(lines);
return lines;
}
}
$(function(){
$('#button').click(function(){
text_split($('#textarea'), 25);
});
});
The html form for the demo:
<textarea id="textarea" rows="10" cols="80">I am totally unappreciated in my time. You can run this whole park from this room with minimal staff for up to 3 days. You think that kind of automation is easy? Or cheap? You know anybody who can network 8 connection machines and debug 2 million lines of code for what I bid for this job? Because if he can I'd like to see him try.</textarea>
<button id="button">demo</button>

Example for 25 characters max, you can use this pattern:
/\S[\s\S]{0,23}\S(?=\s|$)/g
demo
code example:
var text = " I am totally unappreciated in my time. You can run this whole park from this room with minimal staff for up to 3 days. You think that kind of automation is easy? Or cheap? You know anybody who can network 8 connection machines and debug 2 million lines of code for what I bid for this job? Because if he can I'd like to see him try.";
var myRe = /\S[\s\S]{0,23}\S(?=\s|$)/g;
var m;
var result = new Array();
while ((m = myRe.exec(text)) !== null) {
result.push(m[0]);
}
console.log(result);
Note: if you need to choose dynamically the max size, you must use the alternative syntax to define your RegExp object:
var n = 25;
var myRe = new RegExp("\\S[\\s\\S]{0," + (n-2) + "}\\S(?=\\s|$)", "g");
pattern details:
\S # a non-space character (it is obviously preceded by a space
# or the start of the string since the previous match
# ends before a space)
[\s\S]{0,23} # between 0 or 23 characters
\S(?=\s|$) # a non-space character followed by a space or the end of the string
Note that (?=\s|$) can be replaced with (?!\S).

Google Script maker to format a spreadsheet

This is my first time posting so bear with me if I leave any crucial details out.
Anyway, to summarize the problem: I have been trying to get a script to work on Google Script maker to format a spreadsheet which is hooked up to a form, to go straight to my email.
So basically User Form --> Spreadsheet --> My Email
The questions are pretty Standard:
What's the problem?
Where are you located
However the one question I'd like to use is "What is the priority of this problem?" High or low. I have it under multiple choice format so its a simple choice.
Psuedocode for what I want:
if (priority = low) put #priority low onto the email
Simple enough, however I can't seem to get it to work, here's my code:
function sendFormByEmail(e)
{
// Remember to replace XYZ with your own email address
var email = "email";
var subject = "Help Desk Form Submitted";
var s = SpreadsheetApp.getActiveSheet();
var headers = s.getRange(1,1,1,s.getLastColumn()).getValues()[0];
var message = "";
var priority = "";
if(message.indexOf("What is the priority of this problem? = Low")){
priority += "#priority low";
}
else
priority == "GFHHFFHAHFH ";
for(var i in headers){
message += headers[i] + ' = \t \t'+ e.namedValues[headers[i]].toString() + "\n\n"; }
if (message.indexOf("What is the priority of this problem? = Low"))
message += "This is a test";
else
message += "This is not a test";
MailApp.sendEmail(email, subject, message);
}

Let's look at the first instance of if(message.indexOf().... It's got some problems:
A few lines earlier, message was set to an empty string... so you won't find the "priority" string in it.
The if statement is treating the return of indexOf() as a Boolean. However, the return code from .indexOf() is -1 when an item is not found, which is "true-ish". If the string IS found, and is located at the start of the search subject, the return will be 0, which is "false-ish". If it is found at any other location, the value will be `>0', also "true-ish".
In the else, there's a typo. The comparison == should be assignment =, or += if you prefer.
Looking at the surrounding code, this piece looks like it was left-over from a previous version, and can be deleted.
Now look at the second instance.
The message should now be populated. However, the comparison is still using an incorrect Boolean interpretation of .indexOf().
The search string contains a batch of spaces... but the previous code that looped through responses used tabs to separate the "header" from the "value", so the search will always return '-1' (which will be interpreted as true).
There are a couple of other tidy-up items. You probably need:
function sendFormByEmail(e)
{
Logger.log(JSON.stringify(e));
var email = Session.getEffectiveUser().getEmail();
var subject = "Help Desk Form Submitted";
var message = "";
var s = e.range.getSheet(); // Sheet that received form response
var headers = s.getDataRange().getValues()[0];
for (var question in headers) {
message += headers[question] + ' = \t\t' + e.values[question] + '\n\n';
}
// Add text relative to the priority of the reported issue.
if (e.namedValues["What is the priority of this problem?"].toString() === "Low")
message += "This is a test";
else
message += "This is not a test";
Logger.log(message);
MailApp.sendEmail(email, subject, message);
}
PS: You can see why it would be advisable to keep form questions short, like "Priority", and leave the explanatory sentence to be helper text!

Help parsing string (City, State Zip) with JavaScript

I've got a string with the following format:
City, State ZIP
I'd like to get City and State from this string.
How can I do that with JavaScript? edit: note that he doesn't mention he already has the zip code when he gets here, if that helps you in your solution ~~ drachenstern

var address = "San Francisco, CA 94129";
function parseAddress(address) {
// Make sure the address is a string.
if (typeof address !== "string") throw "Address is not a string.";
// Trim the address.
address = address.trim();
// Make an object to contain the data.
var returned = {};
// Find the comma.
var comma = address.indexOf(',');
// Pull out the city.
returned.city = address.slice(0, comma);
// Get everything after the city.
var after = address.substring(comma + 2); // The string after the comma, +2 so that we skip the comma and the space.
// Find the space.
var space = after.lastIndexOf(' ');
// Pull out the state.
returned.state = after.slice(0, space);
// Pull out the zip code.
returned.zip = after.substring(space + 1);
// Return the data.
return returned;
}
address = parseAddress(address);
This is probably better then using regular expressions and String.split(), as it takes into account that the state and city may have spaces.
EDIT: Bug fix: It only included the first word of multi-word state names.
And here's a minified version. :D
function parseAddress(a) {if(typeof a!=="string") throw "Address is not a string.";a=a.trim();var r={},c=a.indexOf(',');r.city=a.slice(0,c);var f=a.substring(c+2),s=f.lastIndexOf(' ');r.state=f.slice(0,s);r.zip=f.substring(s+1);return r;}

There are many ways to do this. Here's a very naive one:
var parts = "City, State ZIP".split(/\s+/); // split on whitespace
var city = parts[0].slice(0, parts[0].length - 1); // remove trailing comma
var state = parts[1];
var zip = parts[2];
Here's one that accounts for the presence of spaces in either the city or state or both:
var parts = "san fran bay, new mex state 666666".split(/\s+|,/),
partition = parts.indexOf(""),
city = parts.slice(0, partition).join(" "),
state = parts.slice(partition + 1, -1).join(" "),
zip = parts.pop();
This last one only works if you're lucky enough to be in an environment that supports destructuring assignment:
var city, statezip, state, zip, parts;
[city, statezip] = "Spaced City, New Mexico ZIP".split(/,\s*/);
parts = statezip.split(/\s+/);
zip = parts.pop();
state = parts.join(" ");
None of these perform any validation, of course.

Ok, since advising regex isn't good, here's my solution. It takes into account cities that have spaces in them, which the other responses here don't seem to do:
var str = "New York, NY 20101";
var cityAndRest = str.split(',');
var city = cityAndRest[0];
var stateAndZip = cityAndRest[1].trim().split(' ');
var state = stateAndZip[0];
var zip = stateAndZip[1];

First assumption: American addresses only.
First find out if the last 5 or the last 10 characters are numeric. A simpler test is to see if the last character is numeric. If so, it's probably got the zip code included. Then a simple test to see if the last 10 contains a space (city #####) or if the last ten include a dash (12345-6789) to figure out if it's a 5 or 5+4 zip. We'll test for a hyphen and no space. (city-du-lac 12345 captures -lac 12345)
Next, all addresses split the city and state by a comma, so we want the last comma. Find the index of the last comma, and split there. I don't know of a city that uses commas in it's name, and I'm sure not gonna let my parser burst on an unknown if I can help it. I do ignore the fact that Washington DC could also be Washington, DC. I figure edge cases are for libraries, not one off scripts.
Lastly, trim everything that remains to remove trailing or leading spaces.
function IsNumeric(n) {
return !isNaN(parseFloat(n)) && isFinite(n);
}
var addr = 'New York City, New York 10101';
//var addr = 'San Bernadino, CA 11111';
function getCityStateZip(addr){
var city; var state;var zip;
city = ''; state = ''; zip = '';
var addrLen = addr.length;
if ( IsNumeric( addr.substring(addrLen - 1) ) ) {
//contains a zipcode - just a sanity check
//get last 10 characters for testing easily
var lastTen = addr.substring( addrLen - 10 );
if ( lastTen.indexOf('-') > 0 && ( lastTen.indexOf(' ') == -1 ) ) {
//found a hyphen and no space (matches our complex rule for zipcodes)
zip = lastTen;
} else {
zip = addr.substring( addrLen - 5 ); //assume a basic 5 zip code
}
}
var zipLen = zip.length;
addrLen = addrLen - zipLen - 1;
addr = addr.substring(0, addrLen ); //remove the chars we just moved into zip
var lastComma = addr.lastIndexOf(',');
if ( lastComma == -1 ) {
//you have a problem, how do you want to handle it?
}
city = addr.substring(0,lastComma); //skip the comma itself, yes?
state = addr.substring(lastComma + 2);
return { 'city':city,'state': state,'zip': zip};
}
getCityStateZip(addr)
IsNumeric js function can be found here Validate decimal numbers in JavaScript - IsNumeric()

Easy way but no validation:
var addrObj={};
parseAddress("Beverly Hills, CA 90210",addrObj);
function parseAddress(address, addrObj){
var arr=address.replace(","," ").split(" ");
addrObj.zip=arr.pop();
addrObj.state=arr.pop();
addrObj.city=arr.join(" ");
}

For this type of thing you might want to use JavaScripts RegEx functions.
Here's some info:
http://www.javascriptkit.com/javatutors/re.shtml