Convert QbyteArray to qreal

JonB

@dziko147
After converting.... what? how?

Tell you what: tell us two things:

Forget about converting anything to float. Show us what the values of the first 5 bytes you receive are. (You say the first byte is a status byte and the next 4 bytes are the value you are interested in, right?)
Tell us what floating point number you know that encodes/conveys/sends. (Please make sure it is not 0!)

KroMignon

@dziko147 said in Convert QbyteArray to qreal:

I get value =0 after converting :/

After converting what?
Please, take time to understand what are the data you want to convert.
I cannot do this for you!

dziko147

@KroMignon
QByteArray rpmdata = QByteArray::fromHex("0102030405"); ---> I want to
convert this .
qDebug() << rpmdata << "this is data" ;
QDataStream stream(rpmdata);
quint8 byte;
float rpmfloat;
stream >> byte; // skip first byte
stream >> rpmfloat; // read float value
qDebug() << rpmfloat << "this is data to float" ;
back.setValue(rpmfloat); ----> I display the value converted here .

JonB

@dziko147
You push raw bytes into a QDataStream. You then try to read that back as, say, a float.

Unless I have misunderstood QDataStream you cannot do that. It is a structured stream. E.g. it will (well, should) contain "markers" for the data and its types.

Experts, I am right here, aren't I (have never used QDataStream)?

UPDATE
For anyone reading this, the expert answers below indicate this is not the case, and my understanding was incorrect.

Christian Ehrlicher

@JonB said in Convert QbyteArray to qreal:

E.g. it will (well, should) contain "markers" for the data and its types.

Not for PODs. For QString e.g. a length field is added. What is added for which build-in Qt type is not documented - QDataStream is not meant to be used with anything else than a QDataStream.

KroMignon

@JonB said in Convert QbyteArray to qreal:

Unless I have misunderstood QDataStream you cannot do that. It is a structured stream. E.g. it will (well, should) contain "markers" for the data and its types.

Of course this is possible, this is no "marker" for serializing/deserializing "base types" like float, double, q(u)int8, q(u)int16, etc.

But "02030405" converted to float is 9.6255135462533111609068148127E-38... almost 0 ;)

KroMignon

@dziko147 said in Convert QbyteArray to qreal:

QByteArray rpmdata = QByteArray::fromHex("0102030405"); ---> I want to
convert this .

What did you expect?
Again, you have to know what you are doing.

Here you are converting this array of bytes { 0x02, 0x03, 0x04, 0x05 } to float.
which is 9.6255135462533111609068148127E-38, almost 0.

If you change the array to something else like "014F030405", which means converting { 0x4f, 0x03, 0x04 , 0x05} to float, you will obtain 2.19807872E9 (almost 2.2 Billion)

==> cf. https://www.binaryconvert.com/convert_float.html

dziko147

@KroMignon I tried this :
QByteArray rpmdata;
rpmdata.resize(4);
rpmdata[0] = 0x43;
rpmdata[1] = 0x96;
rpmdata[2] = 0x7D;
rpmdata[3] = 0xA6;
qDebug() << rpmdata << "this is data" ;

QDataStream stream(rpmdata);

float value;
stream >> value; // read float value
qDebug() << value << "this is data to float" ;

back.setValue(value);

I get :

the 43967DA6 = (300.98163 ) .
this is the the variable

KroMignon

@dziko147 said in Convert QbyteArray to qreal:

the 43967DA6 = (300.98163 ) .

Sorry, my bad, it seems there are something I had forgot about QDataStream: when using float you have to specify floating point precision with stream.setFloatingPointPrecision(QDataStream::SinglePrecision);.

~~The other way is to use qfloat16~~ (WRONG)

QByteArray rpmdata;
rpmdata.resize(4);
rpmdata[0] = 0x43;
rpmdata[1] = 0x96;
rpmdata[2] = 0x7D;
rpmdata[3] = 0xA6;
qDebug() << rpmdata << "this is data" ;

QDataStream stream(rpmdata);

stream.setFloatingPointPrecision(QDataStream::SinglePrecision);
float value;
stream >> value; // read float value
qDebug() << value << "this is data to float" ;

back.setValue(value);

dziko147

@KroMignon thank you it works :) .