Small easy-to-use library for reading and writing csv-files in Qt.
Tested on:
- Ubuntu 14.04 with gcc 4.8.4, Qt 4.8 and higher
- Windows with MinGW, Qt 5.3 and higher
- OS X with clang, Qt 4.8, 5.5, 5.7 and higher
- 1. Quick Example
- 2. Usage
- 3. Requirements
- 4. Build
- 5. Run tests
- 6. Installation
- 7. Examples
- 8. Other
- 9. Creators
#include <QList>
#include <QStringList>
#include <QDir>
#include <QDebug>
#include "qtcsv/stringdata.h"
#include "qtcsv/reader.h"
#include "qtcsv/writer.h"
int main()
{
// prepare data that you want to save to csv-file
QStringList strList;
strList << "one" << "two" << "three";
QtCSV::StringData strData;
strData.addRow(strList);
strData.addEmptyRow();
strData << strList << "this is the last row";
// write to file
QString filePath = QDir::currentPath() + "/test.csv";
QtCSV::Writer::write(filePath, strData);
// read data from file
QList<QStringList> readData = QtCSV::Reader::readToList(filePath);
for ( int i = 0; i < readData.size(); ++i )
{
qDebug() << readData.at(i).join(",");
}
return 0;
}Library could be separated into three parts: Reader, Writer and Containers.
qtcsv library can work with standard Qt containers like QList and QStringList, but also with special ones.
AbstractData is a pure abstract class that provide interface for a container class. Here is how it looks:
class AbstractData
{
public:
explicit AbstractData() {}
virtual ~AbstractData() {}
virtual void addEmptyRow() = 0;
virtual void addRow(const QStringList& values) = 0;
virtual void clear() = 0;
virtual bool isEmpty() const = 0;
virtual int rowCount() const = 0;
virtual QStringList rowValues(const int& row) const = 0;
};As you can see, AbstractData declare virtual functions for adding new rows, getting rows values, clearing all information and so on. Basic stuff for a container class.
If you have said Pure Abstract Class, you must also say Implementation. Don't worry, we have some:
StringData have the same interface as AbstractData class plus some useful functions for inserting rows, removing rows and so on. It stores all data as strings. It is most convenient to use it when information that you want to save in csv-file is represented as strings.
If you store information in different types - integers, floating point values, strings or (almost) anything else (example: [1, 3.14, "check"]) - and you don't want to manually transform each element to string, then you can use QVariant magic. Wrap your data into QVariants and pass it to VariantData class.
Use Reader class to read csv-files / csv-data. Let's see what functions it has.
- Read data to QList<QStringList>
QList<QStringList> readToList(const QString& filePath,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);
QList<QStringList> readToList(QIODevice& ioDevice,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);- filePath - string with absolute path to existent csv-file (example: "/home/user/my-file.csv");
- ioDevice - IO Device that contains csv-formatted data;
- separator (optional) - delimiter symbol, that separate elements in a row (by default it is comma - ",");
- textDelimiter (optional) - text delimiter symbol that enclose each element in a row (by default it is double quoute - ");
- codec (optional) - pointer to the codec object that will be used to read data from the file (by default it is UTF-8 codec).
As a result function will return QList<QStringList> that holds content of the file / IO Device. If all went smooth, list will not be empty and size of it will be equal to the number of rows in csv-data source. Each QStringList will contain elements of the corresponding row.
- Read data to AbstractData-based container
bool readToData(const QString& filePath,
AbstractData& data,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);
bool readToData(QIODevice& ioDevice,
AbstractData& data,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);Second function is a little more advanced and, I hope, a little more useful.
- filePath - string with absolute path to existent csv-file;
- ioDevice - IO Device that contains csv-formatted data;
- data - reference to AbstractData-based class object;
- separator (optional) - delimiter symbol;
- textDelimiter (optional) - text delimiter symbol;
- codec (optional) - pointer to the codec object.
Function will save content of the file / IO Device in data object using virtual function AbstractData::addRow(QStringList). If you pass to the function Reader::readToData() object of class StringData or VariantData, elements of csv-data will be saved in them as strings.
If you are not happy with this fact, you can create your own AbstractData-based container class and implement function addRow(QStringList) in a way you want it.
For example, if you know, that each row of your csv-data contains 3 elements (integer value, floating-point value and string), then in function addRow(QStringList) you can convert first element of QStringList to int, second - to double and save all three elements to some internal container (or do with them whatever you want).
- Read data and process it line-by-line by AbstractProcessor-based processor
bool readToProcessor(const QString& filePath,
AbstractProcessor& processor,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);
bool readToProcessor(QIODevice& ioDevice,
AbstractProcessor& processor,
const QString& separator,
const QString& textDelimiter,
QTextCodec* codec);- filePath - string with absolute path to existent csv-file;
- ioDevice - IO Device that contains csv-formatted data;
- processor - reference to AbstractProcessor-based class object;
- separator (optional) - delimiter symbol;
- textDelimiter (optional) - text delimiter symbol;
- codec (optional) - pointer to the codec object.
This function will read csv-data from file / IO Device line-by-line and pass data to processor object.
AbstractProcessor is an abstract class with two functions:
class AbstractProcessor
{
public:
explicit AbstractProcessor() {}
virtual ~AbstractProcessor() {}
virtual void preProcessRawLine(QString& line) { }
virtual bool processRowElements(const QStringList& elements) = 0;
};When Reader opens a csv-data source (file or IO Device), it starts to read it line by line in a cycle. Each new line Reader first of all pass to processor method preProcessRawLine(QString&). In this method you can edit the line - replace values, remove sensitive information and so on.
After Reader parses elements of the row, it pass them to processor method processRowElements(QStringList). What to do next with these elements - the processor decides. Processor can save elements, filter them, edit and so on. As an example we can consider class ReadToListProcessor (defined in reader.cpp) which simply saves elements into QList.
Use Writer class to write csv-data to files / IO Devices.
bool write(const QString& filePath,
const AbstractData& data,
const QString& separator,
const QString& textDelimiter,
const WriteMode& mode,
const QStringList& header,
const QStringList& footer,
QTextCodec* codec);
bool write(QIODevice& ioDevice,
const AbstractData& data,
const QString& separator,
const QString& textDelimiter,
const QStringList& header,
const QStringList& footer,
QTextCodec* codec);- filePath - string with absolute path to csv-file (new or existent);
- ioDevice - IO Device;
- data - object, that contains information that you want to write to csv-file / IO Device. Writer internally will use QStringList AbstractData::rowValues(int) function to get row values;
- separator (optional) - delimiter symbol (by default it is comma - ",");
- textDelimiter (optional) - text delimiter symbol that enclose each element in a row (by default it is double quoute - ");
- mode (optional) - write mode flag. If it set to WriteMode::REWRITE and csv-file exist, then csv-file will be rewritten. If mode set to WriteMode::APPEND and csv-file exist, then new information will be appended to the end of the file. By default mode set to WriteMode::REWRITE.
- header (optional) - strings that will be written at the beginning of the csv-data, separated with defined separator (empty by default);
- footer (optional) - strings that will be written at the end of the csv-data, separated with defined separator (empty by default);
- codec (optional) - pointer to the codec object that will be used to write data to the file (by default it is UTF-8 codec).
Writer uses CRLF as line ending symbols in accordance with standard. If element of the row contains separator symbol or line ending symbols, such element will be enclosed by text delimiter symbols (or double quoute if you have set empty string as text delimiter symbol).
Qt 4.8 and higher. It is quite possible, that library will be successfully built with older Qt versions (4.7, 4.6, ...).
cd /path/to/folder/with/qtcsv
# Create build directory
mkdir ./build
cd ./build
# Build library. You can choose build type: release or debug
qmake ../qtcsv.pro CONFIG+=[release|debug]
make
# Create build directory for tests
mkdir ./tests
cd ./tests
# Build tests. Besides of setting build type, we set path where linker could find compiled library file.
qmake ../../tests/tests.pro CONFIG+=[release|debug] LIBS+=-L../
makecd /path/to/folder/with/qtcsv
# Create build directory
mkdir ./build
cd ./build
# Build library and tests. See CMakeLists.txt for list of additional options that you can set.
cmake -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTS=ON ..
makeIf you going to build qtcsv library on Windows with MinGW, first of all check that your PATH variable contains paths to Qt and MinGW toolsets. For example, you have installed Qt 5.3 into C:\Qt. Then Qt binaries and libraries will be in folder C:\Qt\5.3\mingw482_32\bin and MinGW binaries will be in C:\Qt\Tools\mingw482_32\bin. Add these paths to the PATH variable so that Windows would know where to look for qmake and make binaries.
cd C:\path\to\folder\with\qtcsv
# Create build directory
mkdir .\build
cd .\build
# Build library. You can choose build type: release or debug. Set DESTDIR to current directory.
qmake ..\qtcsv.pro CONFIG+=[release|debug] DESTDIR=%cd%
mingw32-make
# Create build directory for tests
mkdir .\tests
cd .\tests
# Copy library file into 'tests' directory
copy ..\qtcsv.dll .\
# Build tests
qmake ..\..\tests\tests.pro CONFIG+=[release|debug] DESTDIR=%cd%
mingw32-makecd C:\path\to\folder\with\qtcsv
# Create build directory
mkdir .\build
cd .\build
# Build library and tests. See CMakeLists.txt for list of additional options that you can set.
cmake -G "MinGW Makefiles" -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTS=ON ..
mingw32-makeIf you want to run tests, then use this commands after build of qtcsv:
cd /path/to/folder/with/qtcsv/build/tests
# Set LD_LIBRARY_PATH variable so test binary will know where to search library file.
# Suppose, that library file is located in "build" directory, up a level from current directory.
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$PWD/../
chmod 777 qtcsv_tests
./qtcsv_testscd /path/to/folder/with/qtcsv/build/tests
# Copy library file into "tests" directory
copy ..\*qtcsv.dll .\
qtcsv_tests.exeOn Unix-like OS you can install qtcsv library using this command:
sudo make install
sudo ldconfig -n -v /usr/local/libThis command will copy all compiled files (libqtcsv.so*) from build folder to "/usr/local/lib". Also all headers files will be copied from "./include" folder to "/usr/local/include/".
All installation settings are defined in qtcsv.pro file. See copy_lib_headers and target variables.
For additional information, see Qt documentation about files installation.
If you want to try qtcsv, you can download qtcsv-example project. Don't forget to read README file!
If you want to know more about csv-file format, read RFC 4180 standard.
Also on this page you can find useful tips about how should look proper csv-file.
Author: Antony Cherepanov ([email protected])
Contributors: Patrizio "pbek" Bekerle, Furkan "Furkanzmc" Üzümcü, Martin "schulmar" Schulze, cguentherTUChemnitz, David Jung, Nicu Tofan,
Florian Apolloner, Michael Pollind, Kuba Ober