How to add a table column that does operations in each cell based on the values from another column2?

[HattrickNZ]

asked a similar question here but here it is slightly different and here

if I have a search that gives me something like this:

a b c
1 2 3
4 5 6
7 8 9
how do I add a column d that would do an operation (9/6)-1= 0.5 working on column c. ideally I am looking to get the percentage value be it 0.5, 50 or 50%.

a b c d

1 2 3 (3/0)-1=Na
4 5 6 (6/3)-1= 1
7 8 9 (9/6)-1= 0.5

I am not sure if using delta here is the best way to go. Or should I use head or tail?

here is a test search I am working with
| makeresults count=3 | streamstats count as a | eval a=a+1 | streamstats count as b | eval b=b+10 | streamstats count as c | eval c=c+11 | delta a as a_dif p=1 |

Can anyone advise?

[woodcock · Accepted solution]

Like this:

|noop|stats count AS a|eval a="1:4:7"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
| rename Comment AS "above is for spoofing fake data; below is the solution"
| autoregress c AS d p=1 | fillnull d value=0|eval d=round((c/d)-1,2)

[javiergn]

Try this:

your search here
| streamstats values(c) as d window=1 current=f
| eval d = c/d - 1

You can then use fillnull for those Na values you don't want to display as blank

[woodcock]

I would use | streamstats current=f last(c) as d

[HattrickNZ]

tks. is there a similar parameter as p=1 in the autoregress func below that controls how many values back to use, that can be used in the streamstats dexcribed here? if not autoregress is probably better as it provides that option

[woodcock]

Yes: window=

[HattrickNZ]

hmm not sure.

|noop|stats count AS a|eval a="1:4:7"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
     | rename Comment AS "above is for spoofing fake data; below is the solution" | streamstats values(c) as d current=f window=2

give a different result to this

|noop|stats count AS a|eval a="1:4:7"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
     | rename Comment AS "above is for spoofing fake data; below is the solution"
     | autoregress c AS d p=2 

[woodcock]

I did not say it was identical; it provides similar capability that can be exploited. The window option is actually more flexible but takes more adjustment. You definitely should NOT be using values as indicated by the fact that none of the rest of us are using it. Try this:

|noop|stats count AS a|eval a="1:4:7"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
| rename Comment AS "above is for spoofing fake data; below is the solution"
| streamstats first(c) as d current=f window=2

Yes, I know it isn't exactly the same but the proof of concept is there.

[HattrickNZ]

tks I get it with these 2 examples, which are very similar with the offseting in column d.

|noop|stats count AS a|eval a="1:4:7:8:9:10:11:12:13"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
 | rename Comment AS "above is for spoofing fake data; below is the solution"
 | streamstats first(c) as d current=f window=3

2nd

|noop|stats count AS a|eval a="1:4:7:8:9:10:11:12:13"|makemv delim=":" a|mvexpand a|eval b=a+1|eval c=b+1
 | rename Comment AS "above is for spoofing fake data; below is the solution"
 | autoregress c AS d p=3

[HattrickNZ]

not sure if i have to use streamstats
| makeresults count=3 | streamstats count as a | eval a=a+1 | streamstats count as b | eval b=b+10 | streamstats count as c | eval c=c+11 | delta a as a_dif p=1 | streamstats values(a) as a_vals