# rounding a column vector

20 views (last 30 days)
Ankit Labh on 14 Aug 2018
Commented: Brent F on 23 Jul 2021
I have a column vector with 8000 entries around temp (see array below).
temp=[5 50 150 180 200 220 240 250 260 265 270 272 274 275 275.5 276 277 278 279 279.5 280 280.5 281 281.5 282 284 286 288 290 295 300 310 320]; %temperatures [K]
I did the rounding up to one decimal places. Now I have numbers like 273.9 and 274.1 along with 274. I want to make all such 273.9 and 274.1 to 274, and wherever there is any number with .1 and .9, it should do the floor and ceil with that so that I always get an integer. 275.5 type numbers are the exception.
I tried to use this :
for i= 2:8351
while data(i,2)~=ceil(data(i,2))
if data(i+1,2)== (data(i,2))+0.1
data(i,2)=floor(data(i,2),0);
elseif data(i-1,2)== (data(i,2))-0.1
data(i,2)=ceil(data(i,2),0);
elseif data(i+1,2)== (data(i,2))-0.1
data(i,2)=ceil(data(i,2),0);
elseif data(i-1,2)== (data(i,2))+0.1
data(i,2)=floor(data(i,2),0);
else
data(i,2)=round(data(i,2),1);
end
end
end
Summary: There are 32 temperatures and around 8000 data points near those 32 temperatures. I want to go to one by one in each column and see if it has .1 or .9 and if so, round(or ceil or floor) it to the nearest integer.
Thanks.
Ankit Labh on 14 Aug 2018
Question is not specific to 274 but for all temp.

Stephen23 on 14 Aug 2018
Edited: Stephen23 on 14 Aug 2018
This will round values from (X-1).9 to (X).1 to the integer X, otherwise leave the rounded values alone:
>> vec = 1:0.1:2
vec =
1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 1.90 2.00
>> new = round(vec*10)/10;
>> tmp = round(vec*4);
>> idx = mod(tmp,4)==0;
>> new(idx) = tmp(idx)/4
new =
1.00 1.00 1.20 1.30 1.40 1.50 1.60 1.70 1.80 2.00 2.00
Brent F on 23 Jul 2021
Clever. @Stephen Cobeldick used a multiplier + round (which rounds in both directions) to leave the middle untouched! I sense a general principle here...

M on 14 Aug 2018
Edited: M on 14 Aug 2018
Try this:
temp=[273.9 274.1 274.5];
round(temp*2)/2
ans =
274.0000 274.0000 274.5000
Ankit Labh on 14 Aug 2018
Thank you for your reply. There are 32 temperatures and around 8000 data points near to those 32 temperatures. I want to go to one by one in the column and see if it has .1 or .9 and if so, round(or ceil or floor) it to the nearest integer.