Skip to content

Commit f7e8e83

Browse files
negust-microsoftnegust-microsoft
committed
Adding Iris demo scripts
1 parent a851aa7 commit f7e8e83

File tree

2 files changed

+167
-0
lines changed

2 files changed

+167
-0
lines changed

‎samples/features/machine-learning-services/r/iris/iris-native-scoring.sql‎

Lines changed: 73 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,73 @@
1+
Use sqlr;
2+
GO
3+
/* Step 1: Setup schema */
4+
drop table if exists iris_data, iris_models;
5+
go
6+
create table iris_data (
7+
id int not null identity primary key
8+
, "Sepal.Length" float not null, "Sepal.Width" float not null
9+
, "Petal.Length" float not null, "Petal.Width" float not null
10+
, "Species" varchar(100) null
11+
);
12+
create table iris_models (
13+
model_name varchar(30) not null primary key,
14+
model varbinary(max) not null,
15+
native_model varbinary(max) not null
16+
);
17+
go
18+
19+
/* Step 2: Populate test data from iris dataset in R */
20+
insert into iris_data
21+
("Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width", "Species")
22+
execute sp_execute_external_script
23+
@language = N'R'
24+
, @script = N'iris_data <- iris;'
25+
, @input_data_1 = N''
26+
, @output_data_1_name = N'iris_data';
27+
go
28+
29+
/* Step 3: Create procedure for training model */
30+
create or alter procedure generate_iris_model
31+
(@trained_model varbinary(max) OUTPUT, @native_trained_model varbinary(max) OUTPUT)
32+
as
33+
begin
34+
execute sp_execute_external_script
35+
@language = N'R'
36+
, @script = N'
37+
# Build decision tree model to predict species based on sepal/petal attributes
38+
iris_model <- rxDTree(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, data = iris_rx_data);
39+
40+
# Serialize model to binary format for storage in SQL Server
41+
trained_model <- as.raw(serialize(iris_model, connection=NULL));
42+
43+
# Serialize model to native binary format for scoring using PREDICT function in SQL Server
44+
native_trained_model <- rxSerializeModel(iris_model, realtimeScoringOnly = TRUE)
45+
'
46+
, @input_data_1 = N'
47+
select "Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width", "Species"
48+
from iris_data'
49+
, @input_data_1_name = N'iris_rx_data'
50+
51+
, @params = N'
52+
@trained_model varbinary(max) OUTPUT, @native_trained_model varbinary(max) OUTPUT'
53+
, @trained_model = @trained_model OUTPUT
54+
, @native_trained_model = @native_trained_model OUTPUT;
55+
end;
56+
go
57+
58+
/* Step 3: Train & store a decision tree model that will predict species of flowers */
59+
declare @model varbinary(max), @native_model varbinary(max);
60+
exec generate_iris_model @model OUTPUT, @native_model OUTPUT;
61+
delete from iris_models where model_name = 'iris.dtree';
62+
insert into iris_models (model_name, model, native_model) values('iris.dtree', @model, @native_model);
63+
select model_name, datalength(model)/1024. as model_size_kb, datalength(native_model)/1024. as native_model_size_kb
64+
from iris_models;
65+
go
66+
67+
/* Step 4: Generate predictions using PREDICT function */
68+
declare @native_model varbinary(max) =
69+
(select native_model from iris_models where model_name = 'iris.dtree');
70+
select p.*, d.Species as "Species.Actual", d.id
71+
from PREDICT(MODEL = @native_model, DATA = dbo.iris_data as d)
72+
with(setosa_Pred float, versicolor_Pred float, virginica_Pred float) as p;
73+

‎samples/features/machine-learning-services/r/iris/iris.sql‎

Lines changed: 94 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,94 @@
1+
USE sqlr;
2+
GO
3+
/* Step 1: Setup schema */
4+
drop table if exists iris_data, iris_models;
5+
go
6+
create table iris_data (
7+
id int not null identity primary key
8+
, "Sepal.Length" float not null, "Sepal.Width" float not null
9+
, "Petal.Length" float not null, "Petal.Width" float not null
10+
, "Species" varchar(100) null
11+
);
12+
create table iris_models (
13+
model_name varchar(30) not null primary key,
14+
model varbinary(max) not null,
15+
native_model varbinary(max) not null
16+
);
17+
go
18+
19+
/* Step 2: Populate test data from iris dataset in R */
20+
insert into iris_data
21+
("Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width", "Species")
22+
execute sp_execute_external_script
23+
@language = N'R'
24+
, @script = N'iris_data <- iris;'
25+
, @input_data_1 = N''
26+
, @output_data_1_name = N'iris_data';
27+
go
28+
29+
/* Step 3: Create procedure for training model */
30+
create or alter procedure generate_iris_model
31+
(@trained_model varbinary(max) OUTPUT, @native_trained_model varbinary(max) OUTPUT)
32+
as
33+
begin
34+
execute sp_execute_external_script
35+
@language = N'R'
36+
, @script = N'
37+
# Build decision tree model to predict species based on sepal/petal attributes
38+
iris_model <- rxDTree(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, data = iris_rx_data);
39+
40+
# Serialize model to binary format for storage in SQL Server
41+
trained_model <- as.raw(serialize(iris_model, connection=NULL));
42+
43+
# Serialize model to native binary format for scoring using PREDICT function in SQL Server
44+
native_trained_model <- rxSerializeModel(iris_model, realtimeScoringOnly = TRUE)
45+
'
46+
, @input_data_1 = N'
47+
select "Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width", "Species"
48+
from iris_data'
49+
, @input_data_1_name = N'iris_rx_data'
50+
51+
, @params = N'
52+
@trained_model varbinary(max) OUTPUT, @native_trained_model varbinary(max) OUTPUT'
53+
, @trained_model = @trained_model OUTPUT
54+
, @native_trained_model = @native_trained_model OUTPUT;
55+
end;
56+
go
57+
58+
/* Step 3: Train & store a decision tree model that will predict species of flowers */
59+
declare @model varbinary(max), @native_model varbinary(max);
60+
exec generate_iris_model @model OUTPUT, @native_model OUTPUT;
61+
delete from iris_models where model_name = 'iris.dtree';
62+
insert into iris_models (model_name, model, native_model) values('iris.dtree', @model, @native_model);
63+
select model_name, datalength(model)/1024. as model_size_kb, datalength(native_model)/1024. as native_model_size_kb
64+
from iris_models;
65+
go
66+
67+
/* Step 4: Create procedure for scoring using the decision tree model */
68+
create or alter procedure predict_iris_species (@model varchar(100))
69+
as
70+
begin
71+
declare @rx_model varbinary(max) = (select model from iris_models where model_name = @model);
72+
-- Predict based on the specified model:
73+
exec sp_execute_external_script
74+
@language = N'R'
75+
, @script = N'
76+
# Unserialize model from SQL Server
77+
irismodel<-unserialize(rx_model);
78+
79+
# Predict species for new data using rxDTree model
80+
OutputDataSet <-rxPredict(irismodel, iris_rx_data, extraVarsToWrite = c("Species", "id"));
81+
'
82+
, @input_data_1 = N'
83+
select id, "Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width", "Species"
84+
from iris_data'
85+
, @input_data_1_name = N'iris_rx_data'
86+
, @params = N'@rx_model varbinary(max)'
87+
, @rx_model = @rx_model
88+
with result sets ( ("setosa_Pred" float, "versicolor_Pred" float, "virginica_Pred" float, "Species.Actual" varchar(100), "id" int));
89+
end;
90+
go
91+
92+
/* Step 5: Test scoring of model */
93+
exec predict_iris_species 'iris.dtree';
94+
go

0 commit comments

Comments
 (0)