In the documentation of the DHT11 is said that it has sampling period of minimum 1 second. When I was using one DHT11 I could read new values exactly every 1 second or even less but on another DHT11 when the sampling period is 1 second I'm getting nan every to two measurements. The problem is gone when the sampling period is at least 1,2 seconds so is it normal that it can be different on different sensors?
Another question is if using 4.7 kOhm resitor instead of 5 kOhm as a pull-up for data line can be a problem?
The datasheet I have says it can take up to 15s to respond for humidity and up to 30s for temperature. Why are you polling it every second? Temp and humidity doesn't change that fast. For any situation where you would need a fast response you should be using a better sensor than the DHT11, which doesn't have good accuracy.
No difference between 4.7K and 5K pullup resistor.
I was taking the response time as the amount of time it takes the humitidy or temperature sensor to respond to a change in humidity/temperature, respectively, not the interval between reads of the sensor.
Most references I've seen use two seconds between samples to be safe. Adafruit's DHT library records the time you read the sensor and returns the previous value if it has been less than two seconds.