Develop Reference

Why does FileStream sometimes ignore invisible characters?

I have two blocks of code that I've tried using for reading data out of a file-stream in C#. My overall goal here is to try and read each line of text into a list of strings, but they are all being read into a single string (when opened with read+write access together)...
I am noticing that the first block of code correctly reads in all of my carriage returns and line-feeds, and the other ignores them. I am not sure what is really going on here. I open up the streams in two different ways, but that shouldn't really matter right? Well, in any case here is the first block of code (that correctly reads-in my white-space characters):
StreamReader sr = null;
StreamWriter sw = null;
FileStream fs = null;
List<string> content = new List<string>();
List<string> actual = new List<string>();
string line = string.Empty;
// first, open up the file for reading
fs = File.OpenRead(path);
sr = new StreamReader(fs);
// read-in the entire file line-by-line
while(!string.IsNullOrEmpty((line = sr.ReadLine())))
{
content.Add(line);
}
sr.Close();
Now, here is the block of code that ignores all of the white-space characters (i.e. line-feed, carriage-return) and reads my entire file in one line.
StreamReader sr = null;
StreamWriter sw = null;
FileStream fs = null;
List<string> content = new List<string>();
List<string> actual = new List<string>();
string line = string.Empty;
// first, open up the file for reading/writing
fs = File.Open(path, FileMode.Open);
sr = new StreamReader(fs);
// read-in the entire file line-by-line
while(!string.IsNullOrEmpty((line = sr.ReadLine())))
{
content.Add(line);
}
sr.Close();
Why does Open cause all data to be read as a single line, and OpenRead works correctly (reads data as multiple lines)?
UPDATE 1
I have been asked to provide the text of the file that reproduces the problem. So here it is below (make sure that CR+LF is at the end of each line!! I am not sure if that will get pasted here!)
;$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
;$$$$$$$$$ $$$$$$$
;$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
;
;
;
UPDATE 2
An exact block of code that reproduces the problem (using the text above for the file). In this case I am actually seeing the problem WITHOUT trying Open and only using OpenRead.
StreamReader sr = null;
StreamWriter sw = null;
FileStream fs = null;
List<string> content = new List<string>();
List<string> actual = new List<string>();
string line = string.Empty;
try
{
// first, open up the file for reading/writing
fs = File.OpenRead(path);
sr = new StreamReader(fs);
// read-in the entire file line-by-line
while(!string.IsNullOrEmpty((line = sr.ReadLine())))
{
content.Add(line);
}
sr.Close();
// now, erase the contents of the file
File.WriteAllText(path, string.Empty);
// make sure that the contents of the file have been erased
fs = File.OpenRead(path);
sr = new StreamReader(fs);
if (!string.IsNullOrEmpty(line = sr.ReadLine()))
{
Trace.WriteLine("Failed: Could not erase the contents of the file.");
Assert.Fail();
}
else
{
Trace.WriteLine("Passed: Successfully erased the contents of the file.");
}
// now, attempt to over-write the contents of the file
fs.Close();
fs = File.OpenWrite(path);
sw = new StreamWriter(fs);
foreach(var l in content)
{
sw.Write(l);
}
// read back the over-written contents of the file
fs.Close();
fs = File.OpenRead(path);
sr = new StreamReader(fs);
while (!string.IsNullOrEmpty((line = sr.ReadLine())))
{
actual.Add(line);
}
// make sure the contents of the file are correct
if(content.SequenceEqual(actual))
{
Trace.WriteLine("Passed: The contents that were over-written are correct!");
}
else
{
Trace.WriteLine("Failed: The contents that were over-written are not correct!");
}
}
finally
{
// close out all the streams
fs.Close();
// finish-up with a message
Trace.WriteLine("Finished running the overwrite-file test.");
}

Your new file generated by
foreach(var l in content)
{
sw.Write(l);
}
does not contain end-of-line characters because end-of-line characters are not included in content.
As #DaveKidder points out in this thread over here, the spec for StreamReader.ReadLine specifically says that the resulting line does not include end of line.
When you do
while(!string.IsNullOrEmpty((line = sr.ReadLine())))
{
content.Add(line);
}
sr.Close();
You are losing end of line characters.

Create and write to a text file inmemory and convert to byte array in one go

How can I create a .csv file implicitly/automatically by using the correct method, add text to that file existing in memory and then convert to in memory data to a byte array?
string path = #"C:\test.txt";
File.WriteAllLines(path, GetLines());
byte[] bytes = System.IO.File.ReadAllBytes(path);
With that approach I create a file always (good), write into it (good) then close it (bad) then open the file again from a path and read it from the hard disc (bad)
How can I improve that?
UPDATE
One nearly good approach would be:
using (var fs = new FileStream(#"C:\test.csv", FileMode.Create, FileAccess.ReadWrite))
{
using (var memoryStream = new MemoryStream())
{
fs.CopyTo(memoryStream );
return memoryStream .ToArray();
}
}
but I am not able to write text into that filestream... just bytes...
UPDATE 2
using (var fs = File.Create(#"C:\temp\test.csv"))
{
using (var sw = new StreamWriter(fs, Encoding.Default))
{
using (var ms = new MemoryStream())
{
String message = "Message is the correct ääüö Pi(\u03a0), and Sigma (\u03a3).";
sw.Write(message);
sw.Flush();
fs.CopyTo(ms);
return ms.ToArray();
}
}
}
The string message is not persisted to the test.csv file. Anyone knows why?

Write text into Memory Stream.
byte[] bytes = null;
using (var ms = new MemoryStream())
{
using(TextWriter tw = new StreamWriter(ms)){
tw.Write("blabla");
tw.Flush();
ms.Position = 0;
bytes = ms.ToArray();
}
}
UPDATE
Use file stream Directly and write to File
using (var fs = new FileStream(#"C:\ed\test.csv", FileMode.Create, FileAccess.ReadWrite))
{
using (TextWriter tw = new StreamWriter(fs))
{
tw.Write("blabla");
tw.Flush();
}
}

You can get a byte array from a string using encoding:
Encoding.ASCII.GetBytes(aString);
Or
Encoding.UTF8.GetBytes(aString);
But I don't know why you would want a csv as bytes. You could load the entire file to a string, add to it and then save it:
string content;
using (var reader = new StreamReader(filename))
{
content = reader.ReadToEnd();
}
content += "x,y,z";
using (var writer = new StreamWriter(filename))
{
writer.Write(content);
}
Update: Create a csv in memory and pass back as bytes:
var stringBuilder = new StringBuilder();
foreach(var line in GetLines())
{
stringBuilder.AppendLine(line);
}
return Encoding.ASCII.GetBytes(stringBuilder.ToString());

Decompressing a zipfile into memory stream - C#

I have written code to store the encoded string of zip file into temp path and now I want to store the encoded zipfile string to memorystream instead of temp path. Can someone please help me how to read the stream and pass it as a string to ZipFile class...I am using DOTNETZIP library to unpack password protested file.
Please see below my code.
string tempPath = Path.GetTempPath();
foreach (ActivityMimeAttachment a in attachments.Entities)
{
if (a.FileName.EndsWith(".zip", StringComparison.OrdinalIgnoreCase))
{
string strcontent = a.Body;
byte[] filecontent = Convert.FromBase64String(strcontent); // unpack the base-64 to a blob
File.WriteAllBytes(tempPath + a.FileName, filecontent); // Working code creates a zip file
string attachmentfile = tempPath + a.FileName;
using (ZipFile zip = new ZipFile(attachmentfile))
{
foreach (ZipEntry entry in zip.Entries)
{
if ((entry.FileName.EndsWith(".xml", StringComparison.OrdinalIgnoreCase)) ||
(entry.FileName.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase)))
{
entry.ExtractWithPassword(tempPath, "password");
FileStream inFile;
byte[] binaryData;
string file = tempPath + entry.FileName;
inFile = new FileStream(file, FileMode.Open, FileAccess.Read);
binaryData = new Byte[inFile.Length];
long bytesRead = inFile.Read(binaryData, 0,
(int)inFile.Length);
inFile.Close();

You'll want to convert your file content to a memory stream (Stream filestream = new MemoryStream(filecontent)) then use ZipFile.Read(fileStream). Then use a StreamReader to get the contents out as a string. So try something like this (note it's untested):
string myString;
byte[] filecontent = Convert.FromBase64String(strcontent);
using (var filestream = new MemoryStream(filecontent))
{
using (ZipFile zip = ZipFile.Read(filestream))
{
foreach (ZipEntry entry in zip.Entries)
{
if ((entry.FileName.EndsWith(".xml", StringComparison.OrdinalIgnoreCase)) ||
(entry.FileName.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase)))
{
using (var ms = new MemoryStream())
{
entry.ExtractWithPassword(ms, "password");
ms.Position = 0;
var sr = new StreamReader(ms);
myString = sr.ReadToEnd();
}
...
If the results should be a base64 string, do this:
entry.ExtractWithPassword(ms, "password");
ms.Position = 0;
myString = Convert.ToBase64String(ms.ToArray());
You may or may not have to reset the stream position, but it's good practice to.
Now you can use the results as a string without having to write to a file first.

Is there a better way to write read and modify text lines and write them into an output stream?

I'm currently trying to read a file, modify a few placeholders within and then write the file into an output stream. As its the output stream for a page response in aspx.net I'm using the OutputStream.Write method there (the file is an attachment in the end).
Originally I had:
using (FileStream fs = new FileStream(filename, FileMode.Open, FileAccess.Read, FileShare.ReadWrite))
{
while (readBytes < fs.Length)
{
tmpReadBytes = fs.Read(bytes, 0, bytes.Length);
if (tmpReadBytes > 0)
{
readBytes += tmpReadBytes;
page.Response.OutputStream.Write(bytes, 0, tmpReadBytes);
}
}
}
After thinking things over I came up with the following:
foreach(string line in File.ReadLines(filename))
{
string modifiedLine = line.Replace("#PlaceHolder#", "NewValue");
byte[] modifiedByteArray = System.Text.Encoding.UTF8.GetBytes(modifiedLine);
page.Response.OutputStream.Write(modifiedByteArray, 0, modifiedByteArray.length);
}
But it looks inefficient especially with the conversions. So my question is: Is there any better way of doing this?
As note the file itself is not very big, it's an about 3-4 KB sized textfile.

You don't need to handle the bytes your self.
If you know the file is and always will be small,
this.Response.Write(File.ReadAllText("path").Replace("old", "new"));
otherwise
using (var stream = new FileStream("path", FileMode.Open))
{
using (var streamReader = new StreamReader(stream))
{
while (streamReader.Peek() != -1)
{
this.Response.Write(streamReader.ReadLine().Replace("old", "new"));
}
}
}

To get the lines in a string array:
string[] lines = File.ReadAllLines(file);
To alter the lines, use a loop.
for (int i = 0; i < lines.Length; i++)
{
lines[i] = lines[i].Replace("#PlaceHolder#", "NewValue");
}
And to save the new text, first create a string with all the lines.
string output = "";
foreach(string line in lines)
{
output+="\n"+line;
}
And then save the string to the file.
File.WriteAllText(file,output);

C# ZipArchive losing data

I'm trying to copy the contents of one Excel file to another Excel file while replacing a string inside of the file on the copy. It's working for the most part, but the file is losing 27 kb of data. Any suggestions?
public void ReplaceString(string what, string with, string path) {
List < string > doneContents = new List < string > ();
List < string > doneNames = new List < string > ();
using(ZipArchive archive = ZipFile.Open(_path, ZipArchiveMode.Read)) {
int count = archive.Entries.Count;
for (int i = 0; i < count; i++) {
ZipArchiveEntry entry = archive.Entries[i];
using(var entryStream = entry.Open())
using(StreamReader reader = new StreamReader(entryStream)) {
string txt = reader.ReadToEnd();
if (txt.Contains(what)) {
txt = txt.Replace(what, with);
}
doneContents.Add(txt);
string name = entry.FullName;
doneNames.Add(name);
}
}
}
using(MemoryStream zipStream = new MemoryStream()) {
using(ZipArchive newArchive = new ZipArchive(zipStream, ZipArchiveMode.Create, true, Encoding.UTF8)) {
for (int i = 0; i < doneContents.Count; i++) {
int spot = i;
ZipArchiveEntry entry = newArchive.CreateEntry(doneNames[spot]);
using(var entryStream = entry.Open())
using(var sw = new StreamWriter(entryStream)) {
sw.Write(doneContents[spot]);
}
}
}
using(var fileStream = new FileStream(path, FileMode.Create)) {
zipStream.Seek(0, SeekOrigin.Begin);
zipStream.CopyTo(fileStream);
}
}
}
I've used Microsoft's DocumentFormat.OpenXML and Excel Interop, however, they are both lacking in a few main components that I need.
Update:
using(var fileStream = new FileStream(path, FileMode.Create)) {
var wrapper = new StreamWriter(fileStream);
wrapper.AutoFlush = true;
zipStream.Seek(0, SeekOrigin.Begin);
zipStream.CopyTo(wrapper.BaseStream);
wrapper.Flush();
wrapper.Close();
}

Try the process without changing the string and see if the file size is the same. If so then it would seem that your copy is working correctly, however as Marc B suggested, with compression, even a small change can result in a larger change in the overall size.