simpledataparser's Introduction

Simple Data Parser

Simple data parser and filter recognizing boolean expression and value (with one-time variable substitution) comparisons.

Grammar (EBNF)

<expression> ::= <term> {<or> <term>}
<term> ::= <factor> {<and> <factor>}
<factor> ::= <comparison> | (<expression>)
<comparison> ::= <value> <operator> <value>
<value> ::= <variable> | <number>
<operator> ::= > | >= | < | <= | = | !=
<or> ::= OR | or
<and> ::= AND | and

How to use it

Build project
```
./gradlew build
```
Run program with path to data file as an argument
```
java -jar build/libs/sdp-1.0.jar data/iris.data
```

Enter selector to filter rows

 Enter data selector for attributes [sl, sw, pl, pw]:  (sw < 3 OR sw >= 4) AND pw = 0.2

Look at the constructed abstract syntax tree

 AST for entered selector:
 └── AND
     ├── OR
     │   ├── <
     │   │   ├── sw
     │   │   └── 3.0
     │   └── >=
     │       ├── sw
     │       └── 4.0
     └── =
         ├── pw
         └── 0.2

Accept selector or deny and enter new one
```
 Do you accept it? [y/n]:  y
```

Look at the selected rows

 Selected rows:
 {sw=2.9, pw=0.2, sl=4.4, pl=1.4}
 {sw=4.0, pw=0.2, sl=5.8, pl=1.2}
 {sw=4.2, pw=0.2, sl=5.5, pl=1.4}
 
 Result:  3 out of 150 rows selected in 0.028s

Data files

File format

Input files should contain tab-separated values. If first line contains values that start with a letter it is treated as a header and its values as attribute names. Otherwise attribute names are automatically generated as follows: a, b, ..., z, aa, ab, .... All values outside header should be numerical.