My table looks like this:
ID AQ_ATC amountATC
. "A05" 1
123 "A05AA02" 2525
234 "A05AA02" 2525
991 "A05AD39" 190
. "C10" 1
441 "C10AA11" 4330
229 "C10AA22" 3100
. "C05AA" 1
441 "C05AA03" 130
The count for the full 8-character AQ_ATC codes is already correct.
The shorter codes are unique in the table and are substrings of the complete 8-character codes (they represent the first x characters).
What I am looking for is the count of the appearances of the shorter codes throughout the entire table.
For example in this case the resulting table would be
ID AQ_ATC amountATC
. "A05" 2715 <-- 2525 + 190
123 "A05AA02" 2525
234 "A05AA02" 2525
991 "A05AD39" 190
. "C10" 7430 <-- 4330 + 3100
441 "C10AA11" 4330
229 "C10AA22" 3100
. "C05AA" 130 <-- 130
441 "C05AA03" 130
The partial codes do not overlap, by what I mean that if there is "C05" there wont be another partial code "C05A1".
I created the amountATC column using
bysort ATC: egen amountATC = total(AQ_ATC==AQ_ATC)
I attempted recycling the code that I had received yesterday but failed in doing so.
My attempt looks like this:
levelsof AQ_ATC, local(ATCvals)
quietly foreach y in AQ_ATC {
local i = 0
quietly foreach x of local ATCvals {
if strpos(`y', `"`x'"') == 1{
local i = `i'+1
replace amountATC = `i'
}
}
}
My idea was to use a counter "i" and increase it by 1 everytime the an AQ_ATC starts with another AQ_ATC code. Then I write "i" into amountATC and after I iterated over the entire table for my AQ_ATC, I will have an "i"-value that will be equal to the amount of occurences of the substring. Then I reset "i" to 0 and continue with the next AQ_ATC.
At least thats how I intended for it to work, what it did in the end is set all amountATC-values to 1.
I also attempted looking into different egen-functions such as noccur and moss, but my connection keeps timing out when I attempt to install the packages.
It seems as if you come from another language and you insist in using loops when not strictly necessary. Stata does many things without explicit loops, precisely because commands already apply to all observations.
One way is:
clear
set more off
input ///
ID str15 AQ_ATC amountATC
. "A05" 1
123 "A05AA02" 2525
234 "A05AA02" 2525
991 "A05AD39" 190
. "C10" 1
441 "C10AA11" 4330
229 "C10AA22" 3100
. "C05AA" 1
441 "C05AA03" 130
end
*----- what you want -----
sort AQ_ATC ID
gen grou = sum(missing(ID))
bysort grou AQ_ATC: gen tosum = amountATC if _n == 1 & !missing(ID)
by grou: egen s = total(tosum)
replace amountATC = s if missing(ID)
list, sepby(grou)
Edit
With your edit the same principles apply. Below code that adjusts to your change and slightly changes the code (one line less):
*----- what you want -----
sort AQ_ATC
gen grou = sum(missing(ID))
bysort grou: gen s = sum(amountATC) if AQ_ATC != AQ_ATC[_n+1] & !missing(ID)
by grou: replace amountATC = s[_N] if missing(ID)
More efficient should be:
<snip>
bysort grou: gen s = sum(amountATC) if AQ_ATC != AQ_ATC[_n+1]
by grou: replace amountATC = s[_N] - 1 if missing(ID)
Some comments:
sort is a very handy command. If you sort the data by AQ_ATC they are arranged in such a way that the short (sub)strings are placed before corresponding long strings.
The by: prefix is fundamental and very helpful, and I noticed you can use it after defining appropriate groups. I created the groups taking advantage of the fact that all short (sub)strings have a missing(ID).
Then (by the groups just defined) you only want to add up one value (observation) per amountATC. That's what the condition if AQ_ATC != AQ_ATC[_n+1] does.
Finally, replace back into your original variable. I would usually generate a copy and work with that, so my original variable doesn't suffer.
An excellent read for the by: prefix is Speaking Stata: How to move step by: step, by Nick Cox.
Edit2
Yet another slightly different way:
*----- what you want -----
sort AQ_ATC
gen grou = sum(missing(ID))
egen t = tag(grou AQ_ATC)
bysort grou: gen s = sum(amountATC * t)
by grou: replace amountATC = s[_N] - 1 if missing(ID)
Related
while working on the exercise 2.2 of "programming in Lua 4" I do have to create a function to built all permutations of the numbers 1-8. I decided to use Heaps algorithm und made the following script. I´m testing with numbers 1-3.
In the function I store the permutations as tables {1,2,3} {2,1,3} and so on into local "a" and add them to global "perm". But something runs wrong and at the end of the recursions I get the same permutation on all slots. I can´t figure it out. Please help.
function generateperm (k,a)
if k == 1 then
perm[#perm + 1] = a -- adds recent permutation to table
io.write(table.unpack(a)) -- debug print. it shows last added one
io.write("\n") -- so I can see the algorithm works fine
else
for i=1,k do
generateperm(k-1,a)
if k % 2 == 0 then -- builts a permutation
a[i],a[k] = a[k],a[i]
else
a[1],a[k] = a[k],a[1]
end
end
end
end
--
perm = {}
generateperm(3,{1,2,3}) -- start
--
for k,v in ipairs (perm) do -- prints all stored permutations
for k,v in ipairs(perm[k]) do -- but it´s 6 times {1,2,3}
io.write(v)
end
io.write("\n")
end
debug print:
123
213
312
132
231
321
123
123
123
123
123
123
working on an economic optimization problem with pyomo, I would like to add a constraint to prevent the product of the commodity quantity and its price to go below zero (<0), avoiding a negative revenue. It appears that all the data are in a dataframe and I can't setup a constraint like:
def positive_revenue(model, t)
return model.P * model.C >=0
model.positive_rev = Constraint(model.T, rule=positive_revenue)
The system returns the error that the price is a scalar and it cannot process it. Indeed the price is set as such in the model:
model.T = Set(doc='quarter of year', initialize=df.quarter.tolist(), ordered=True)
model.P = Param(initialize=df.price.tolist(), doc='Price for each quarter')
##while the commodity is:
model.C = Var(model.T, domain=NonNegativeReals)
I just would like to apply that for each timestep (quarter of hour here) that:
price(t) * model.C(t) >=0
Can someone help me to spot the issue ? Thanks
Here are more information:
df dataframe:
df time_stamp price Status imbalance
quarter
0 2021-01-01 00:00:00 64.84 Final 16
1 2021-01-01 00:15:00 13.96 Final 38
2 2021-01-01 00:30:00 12.40 Final 46
index = quarter from 0 till 35049, so it is ok
Here is the df.info()
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 time_stamp 35040 non-null datetime64[ns]
1 price 35040 non-null float64
2 Status 35040 non-null object
3 imbalance 35040 non-null int64
I modified the to_list() > to_dict() in model.T but still facing the same issue:
KeyError: "Cannot treat the scalar component 'P' as an indexed component" at the time model.T is defined in the model parameter, set and variables.
Here is the constraint where the system issues the error:
def revenue_positive(model,t):
for t in model.T:
return (model.C[t] * model.P[t]) >= 0
model.positive_revenue = Constraint(model.T,rule=revenue_positive)
Can't figure it out...any idea ?
UPDATE
Model works after dropping an unfortunate 'quarter' column somewhere...after I renamed the index as quarter.
It runs but i still get negative revenues, so the constraints seems not working at present, here is how it is written:
def revenue_positive(model,t):
for t in model.T:
return (model.C[t] * model.P[t]) >= 0
model.positive_revenue = Constraint(model.T,rule=revenue_positive)
What am I missing here ? Thanks for help, just beginning
Welcome to the site.
The problem you appear to be having is that you are not building your model parameter model.P as an indexed component. I believe you likely want it to be indexed by your set model.T.
When you make indexed params in pyomo you need to initialize it with some key:value pairing, like a python dictionary. You can make that from your data frame by re-indexing your data frame so that the quarter labels are the index values.
Caution: The construction you have for model.T and this assume there are no duplicates in the quarter names.
If you have duplicates (or get a warning) then you'll need to do something else. If the quarter labels are unique you can do this:
import pandas as pd
import pyomo.environ as pyo
df = pd.DataFrame({'qtr':['Q5', 'Q6', 'Q7'], 'price':[12.80, 11.50, 8.12]})
df.set_index('qtr', inplace=True)
print(df)
m = pyo.ConcreteModel()
m.T = pyo.Set(initialize=df.index.to_list())
m.price = pyo.Param(m.T, initialize=df['price'].to_dict())
m.pprint()
which should get you:
price
qtr
Q5 12.80
Q6 11.50
Q7 8.12
1 Set Declarations
T : Size=1, Index=None, Ordered=Insertion
Key : Dimen : Domain : Size : Members
None : 1 : Any : 3 : {'Q5', 'Q6', 'Q7'}
1 Param Declarations
price : Size=3, Index=T, Domain=Any, Default=None, Mutable=False
Key : Value
Q5 : 12.8
Q6 : 11.5
Q7 : 8.12
2 Declarations: T price
edit for clarity...
NOTE:
The first argument when you create a pyomo parameter is the indexing set. If this is not provided, pyomo assumes that it is a scalar. You are missing the set as shown in my example and highlighted with arrow here: :)
|
|
|
V
m.price = pyo.Param(m.T, initialize=df['price'].to_dict())
Also note, you will need to initialize model.P with a dictionary as I have in the example, not a list.
I am examining prescription patterns within a large EHR dataset. The data is structured so that we are given several key bits of information, such as patient_num, encounter_num, ordering_date, medication, age_event (age at event) etc. Example below:
Patient_num enc_num ordering_date medication age_event
1111 888888 07NOV2008 Wellbutrin 48
1111 876578 11MAY2011 Bupropion 50
2222 999999 08DEC2009 Amitriptyline 32
2222 999999 08DEC2009 Escitalopram 32
3333 656463 12APR2007 Imipramine 44
3333 643211 21DEC2008 Zoloft 45
3333 543213 02FEB2009 Fluoxetine 45
Currently I have the dataset sorted by patient_id then by ordering_date so that I can see what each individual was prescribed during their encounters in a longitudinal fashion. For now, I am most concerned with the prescription(s) that were made during their first visit. I wrote some code to count the number of prescriptions and had originally restricted later analyses to RX = 1, but as we can see, that doesn't work for people with multiple scripts on the same encounter (Patient 2222).
data pt_meds_;
set pt_meds;
by patient_num;
if first.patient_num then RX = 1;
else RX + 1;
run;
Patient_num enc_num ordering_date medication age_event RX
1111 888888 07NOV2008 Wellbutrin 48 1
1111 876578 11MAY2011 Bupropion 50 2
2222 999999 08DEC2009 Amitriptyline 32 1
2222 999999 08DEC2009 Escitalopram 32 2
3333 656463 12APR2007 Imipramine 44 1
3333 643211 21DEC2008 Zoloft 45 2
3333 543213 02FEB2009 Fluoxetine 45 3
I think it would be more appropriate to recode the encounter numbers into a new variable so that they reflect a style similar to the RX variable. Where each encounter is listed 1-n, and the number will repeat if multiple scripts are made in the same encounter. Such as below:
Patient_num enc_num ordering_date medication age_event RX Enc_
1111 888888 07NOV2008 Wellbutrin 48 1 1
1111 876578 11MAY2011 Bupropion 50 2 2
2222 999999 08DEC2009 Amitriptyline 32 1 1
2222 999999 08DEC2009 Escitalopram 32 2 1
3333 656463 12APR2007 Imipramine 44 1 1
3333 643211 21DEC2008 Zoloft 45 2 2
3333 543213 02FEB2009 Fluoxetine 45 3 3
From what I have seen, this could be possible with a variant of the above code using 2 BY groups (patient_num & enc_num), but I can't seem to get it. I think the first. / last. codes require sorting, but if I am to sort by enc_num, they won't be in chronological order because the encounter numbers are generated by the system and depend on all other encounters going in at that time.
I tried to do the following code (using ordering_date instead because its already sorted properly) but everything under Enc_ is printed as a 1. I'm sure my logic is all wrong. Any thoughts?
data pt_meds_test;
set pt_meds_;
by patient_num ordering_date;
if first.patient_num;
if first.ordering_date then enc_ = 1;
else enc_ + 1;
run;
First
.First/.Last flags doesn't require sorting if data is properly ordered or you use NOTSORTED in your BY statement. If your variable in BY statement is not properly ordered then BY statment will throw error and stop executing when encounter deviations. Like this:
data class;
set sashelp.class;
by age;
first = first.age;
last = last.age;
run;
ERROR: BY variables are not properly sorted on data set SASHELP.CLASS.
Name=Alfred Sex=M Age=14 Height=69 Weight=112.5 FIRST.Age=1 LAST.Age=1 first=. last=. _ERROR_=1 _N_=1
NOTE: The SAS System stopped processing this step because of errors.
NOTE: There were 2 observations read from the data set SASHELP.CLASS.
Try this code to see how exacly .first/.last flags works:
data pt_meds_test;
set pt_meds_;
by patient_num ordering_date;
fp = first.patient_num;
lp = last.patient_num;
fo = first.ordering_date;
lo = last.ordering_date;
run;
Second
Those condidions works differently than you think:
if expression;
If expression is true then continue with next instructions after if.
Otherwise return to begining of data step (no implicit output). This also implies your observation is not retained in the output.
In most cases if without then is equivalent to where. However
whereworks faster but it is limited to variables that comes from data set you are reading
if can be used with any type of expression including calculated fields
More info:: IF
Statement, Subsetting
Third
I think lag() function can be your answear.
data pt_meds_test;
set pt_meds_;
by patient_num;
retain enc_;
prev_patient_num = lag(patient_num);
prev_ordering_date = lag(ordering_date);
if first.patient_num then enc_ = 1;
else if patient_num = prev_patient_num and ordering_date ne prev_ordering_date then enc_ + 1;
end;
run;
With lag() function you can look what was the value of vairalbe on the previos observation and compare it with current one later.
But be carefull. lag() doesn't look for variable value from previous observation. It takes vale of variable and stores it in a FIFO queue with size of 1. On next call it retrives stored value from queue and put new value there.
More info: LAG Function
I'm not sure if this hurts the rest of your analysis, but what about just
proc freq data=pt_meds noprint;
tables patient_num ordering_date / out=pt_meds_freq;
run;
data pt_meds_freq2;
set pt_meds_freq;
by patient_num ordering_date;
if first.patient_num;
run;
I am new to R and stuck up in a very naive thing. I am getting 'NA' values in count array after executing following code:
i=1
j=2
l=1
count=0
while(j<length(positions)){
a=positions[i]
b=positions[j]
for(k in a:b){
if(y$feature[k]==x$feature[l]){
count[l]=count[l]+1
}
}
i=i+2
j=j+2
l=l+1
}
For reference, y and x data frames are as follows:
y data frame
positions id feature
1 1 45128
2 1 28901
3 1 48902
. .
. .
. .
. .
2344 1 45579
2345 2 37689
2346 2 45547
. .
. .
5677 2 12339
5678 3 98034
5679
.
.
x dataframe :
id feature
1 28901
2 23498
3 98906
. .
. .
. .
I have inserted the positions in the position array, at the point where new id starts and where it ends
positions is an array consisting of [1,2344,2345,5677,5678,7390,7391,...]. I am incrementing the for loop as elements in position array, i being 1,3,5... j being 2,4,6... If y$feature and x$feature match I increment count[l]
So first feature of x is compared with all features in y with id=1, second feature in x is compared with all features in y with id=2 and so on. When they match, count[l] is incremented. i and j are incremented twice, to make them start with correct positions. *But I just get a valid answer for count[1], rest all values are NA.
Please tell a reason why this happens and a valid way to do this using the loops.
It's because you are trying to add a nonexistent value count[l] to 1. You start out with count<-0, so count is of length one. There is no count[2], so a reference to count[2] returns NA. Then (assuming l = 2 in your loop), NA + l returns NA.
If you initialize count<-rep(0,length(positions)) this particular problem will go away.
Meanwhile, you can vectorize your operations quite a lot. I believe you can replace the k-loop with
count[l] <- sum(y$feature[a:b]==x$feature[l])
for one example.
Currently I'm struggling with an AWK problem that I haven't been able to solve yet. I have one huge file (30GB) with genomic data that holds a list with positions (declared in col 1 and 2) and a second list that holds a number of ranges (declared in col 3, 4 and 5). I want to extract all lines in the first file where the position falls within the range declared in the seconds file. As the position is only unique within a certain chromosome (chr) first it has to be tested if the chr's are identical (ie. col1 in file 1 matches col3 in file2)
file 1
chromosome position another....hundred.....columns
chr1 816 .....
chr1 991 .....
chr2 816 .....
chr2 880 .....
chr2 18768 .....
...
chr22 9736286 .....
file 2
name identifier chromosome start end
GENE1 ucsc.86 chr1 800 900
GENE2 ucsc.45 chr2 700 1700
GENE3 ucsc.46 chr2 18000 19000
expected output
chromosome position another....hundred.....columns
chr1 816 .....
chr2 816 .....
chr2 880 .....
chr2 18768 .....
A summery of what I intend to do in (half coded):
(if $1(in file 1) matches $3(in file 2){ ##test if in the correct chr
if ($2(in file 1) >= $4 && =< $5 (in file 2){ ##test if pos is in the range
print $0 (in file 1) ##if so print the row from file1
}
}
I kind if understand how to solve this problem by putting file1 in an array and using position as the index but then I still have a problem with the chr and besides that file1 is way to big to put in an array (although I have 128GB of RAM). I've tried some things with multi-dimensional arrays but couldn't really figure out how to do that either.
Thanks a lot for all your help.
Update 8/5/14
Added a third line in file 2 containing another range in the same chrom. as on the second line. This line is skipped in the script below.
It'd be something like this, untested:
awk '
NR==FNR{ start[$3] = $4; end[$3] = $5; next }
(FNR==1) || ( ($1 in start) && ($2 >= start[$1]) && ($2 <= end[$1]) )
' file2 file1
The change in your data set actually modified the question greatly. You introduced an element which was used as a key and since keys have to be unique it got overwritten.
For your data set, you are better off making composite keys. Something like:
awk '
NR==FNR{ range[$3,$4,$5]; next }
FNR==1
{
for(x in range) {
split(x, check, SUBSEP);
if($1==check[1] && $2>=check[2] && $2<=check[3]) print $0
}
}
' file2 file1
chromosome position another....hundred.....columns
chr1 816 .....
chr2 816 .....
chr2 880 .....
chr2 18768