Thank you so much you explained in detail.
I could make sense that In general though one can rely upon 15 significant digits to the right of the decimal point.
If the results value is very small, I use that 15 significant digits to the right of the decimal point, such as 0.000000000000001.
Original Message:
Sent: Mon June 24, 2024 10:59 AM
From: Curtis Browning
Subject: I would like to know the result displayed with the maximum number of digits after the decimal point.
Hi Nakayama-san,
If I may add to the excellent information provided by my friend and former colleague Jon K Peck:
Jon is absolutely correct that IBM SPSS Statistics is constrained by the limitations of the 8-byte double-precision values that we use. The IEEE 754 double-precision format is a powerful and somewhat flexible definition for how floating-point values can be represented using 8-byte storage units on modern computer hardware.
Therefore there is not just one hard and fast rule for how much precision can be accurately represented to the right of a decimal point. It depends upon such factors as the magnitude of the value to the left of the decimal point, and also the sign of the value (positive or negative). In general though one can rely upon 15 significant digits to the right of the decimal point. Anything smaller than that will not always be guaranteed to provide exactly the same value when converted to a string and then back to an IEEE double.
See here for an overview of this topic. In particular note the section entitled "IEEE 754 double-precision binary floating-point format: binary64": https://en.wikipedia.org/wiki/Double-precision_floating-point_format.
Best regards,
Curtis Browning
------------------------------
Curtis Browning
SPSS Statistics Architect
Original Message:
Sent: Fri June 21, 2024 03:28 AM
From: NAKAYAMA Kyoko
Subject: I would like to know the result displayed with the maximum number of digits after the decimal point.
Thank you so much you replied.
I understood using scientific notation with 15 decimals as the significant figures.
>>However, you can double click the table and change the displayed cell format as you wish.--
I tried change the displayed cell format as the maximum number of digits after the decimal point.
It was available to change up to 36digits.
In other words, is it possible to display the full precision value after the decimal point (very small value) like "0.0000000000000000000000000000000001"?
Thanks so much,
For example, is it available the most exact detail in the display like "0.000000000000000000000000000000000001".
------------------------------
NAKAYAMA Kyoko
Original Message:
Sent: Thu June 20, 2024 09:33 AM
From: Jon Peck
Subject: I would like to know the result displayed with the maximum number of digits after the decimal point.
What matters in the output is significant figures, not the number of decimal places. If you want to see the most exact detail in the display, use scientific notation with 15 decimals, which is the maximum due to the nature of floating point hardware.
--
Original Message:
Sent: 6/20/2024 12:46:00 AM
From: NAKAYAMA Kyoko
Subject: RE: I would like to know the result displayed with the maximum number of digits after the decimal point.
Thank you so much you replied.
I would you like to make sure about your message.
Do you mean that full precision value even if they are rounded for display purposes so number of digits that can be changed in cell formatting is 36 digits?
Thanks so much,
------------------------------
NAKAYAMA Kyoko
Original Message:
Sent: Mon June 17, 2024 09:31 PM
From: Jon Peck
Subject: I would like to know the result displayed with the maximum number of digits after the decimal point.
You should understand that values in pivot tables and transformations are computed to the full precision value even if they are rounded for display purposes. You can change the cell format display to show more digits either interactively in the pivot table or by creating a tablelook with a differnt cell format and setting that as your default.
------------------------------
Jon Peck
Original Message:
Sent: Mon June 17, 2024 02:11 AM
From: NAKAYAMA Kyoko
Subject: I would like to know the result displayed with the maximum number of digits after the decimal point.
Could you tell me the answer following questions.
①What is the maximum number of decimal places in the analysis result?
②How are analysis results that are extremely close to 0 displayed?
I would like to know the result displayed with the maximum number of digits after the decimal point.
------------------------------
中山 京子
------------------------------