I wrote one function that return a pair of QDateTime
as observable, like this one:
rxcpp::observable<std::tuple<QDateTime, QDateTime>> experimentOne(const QDateTimeAxis * const axis
{
return rxcpp::observable<>::create<std::tuple<QDateTime, QDateTime>>(
[axis](rxcpp::subscriber<std::tuple<QDateTime, QDateTime>> s) {
auto rangeCallback = [s](QDateTime minv, QDateTime maxv) {
if (s.is_subscribed()) {
// send to the subscriber
s.on_next(std::make_tuple<QDateTime, QDateTime>(std::move(minv), std::move(maxv)));
}
};
QObject::connect(axis, &QDateTimeAxis::rangeChanged, rangeCallback);
});
}
So with this function I can subscribe to change in the date range on a axis of a QChart
.
I wrote also another function that, given two dates, returns an observable with values coming from a sqlite db, like the following
rxcpp::observable<std::tuple<double, double>> Database::getValueRange(const std::string& table, unsigned long start, unsigned long end)
{
return rxcpp::observable<>::create<std::tuple<double, double>>(
[this, table, start, end](rxcpp::subscriber<std::tuple<double, double>> s) {
// get the prepared statement for the query 1, i.e. ohlcv values
// within a date range
sqlite3_stmt *stmt = this->m_query3_stms[table].get();
// bind first parameter, the start timestamp
int rc = sqlite3_bind_int64(stmt, 1, start);
checkSqliteCode(rc, m_db.get());
// bind the second parameter, the end timestamp
rc = sqlite3_bind_int64(stmt, 2, end);
checkSqliteCode(rc, m_db.get());
// step through the query results
while ( sqlite3_step(stmt)==SQLITE_ROW && s.is_subscribed() ) {
// extract name values from the current result row
float minv = sqlite3_column_double(stmt, 0);
float maxv = sqlite3_column_double(stmt, 1);
// send to the subscriber
s.on_next(std::make_tuple<double, double>(minv, maxv));
}
// reset the statement for reuse
sqlite3_reset(stmt);
// send complete to the subscriber
s.on_completed();
});
}
How can I pipe the values from the first functions (two dates) as input to the second function in a idiomatic form in RxCpp? In a way at the end of the pipeline I can subscribe to the values coming from the db based on the input dates?
The canonical way to create a new value range for each new pair of date values is to use map followed by one of the flattening operators
auto valueRanges = experimentOne(/*params*/).
map(rxcpp::util::apply_to([](QDateTime d1, QDateTime d2){
return getValueRange(/*params*/).
map(rxcpp::util::apply_to([=](double db1, double db2){
return std::make_tuple(d1, d2, db1, db2);
}));
})).
/* switch_on_next() or merge() or concat() */
/* this will produce std::tuple< QDateTime, QDateTime, double, double>
switch_on_next
will cancel the previous value range and start the new value range.merge
will produce all the value ranges as soon as possible.concat
will produce the values ranges one at a time, in order.in the case that the values ranges are run on different threads a thread-safe coordination must be passed to merge
so that the value ranges are interleaved safely.
To select out a particular range, use filter()
. If you want to be able to split out the ranges into separate expressions, use publish()
to share the ranges first.
auto hotValueRanges = valueRanges.
publish().ref_count();
auto aDateRange = hotValueRanges.
filter(rxcpp::util::apply_to([](QDateTime d1, QDateTime d2, double, double){
return isADesiredDate(d1, d2);
})).
subscribe(/*use the range*/);
auto anotherDateRange = hotValueRanges.
filter(rxcpp::util::apply_to([](QDateTime d1, QDateTime d2, double, double){
return isAnotherDesiredDate(d1, d2);
})).
subscribe(/*use the range*/);