Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/ahmedalhulaibi/flatfile

A library to unmarshal structured data read from a flat-file (text file) into a struct using struct tags
https://github.com/ahmedalhulaibi/flatfile

converter flat-file go golang unmarshaller

Last synced: 22 days ago
JSON representation

A library to unmarshal structured data read from a flat-file (text file) into a struct using struct tags

Awesome Lists containing this project

README

        

# flatfile

[![Pipeline](https://gitlab.com/ahmedalhulaibi/flatfile/badges/master/pipeline.svg)](https://gitlab.com/ahmedalhulaibi/flatfile/pipelines)[![codecov](https://codecov.io/gl/ahmedalhulaibi/flatfile/branch/master/graph/badge.svg)](https://codecov.io/gl/ahmedalhulaibi/flatfile)

This package provides a utility to read a fixed-width record from a [flat-file database](https://en.wikipedia.org/wiki/Flat-file_database) or a text file. The intent is to eliminate boilerplate code for reading data from a flat file and mapping it to the fields in a struct.

This package allows you to define your record layout mapping using struct tags.

This packge provides a method `Unmarshal` which will convert a record (slice of bytes) into a struct.

# Usage

## Download

Use your favourite dependency tool to pull the code. dep, go mod, etc.

`go get github.com/ahmedalhulaibi/flatfile`

## Examples

Please refer to the [examples](https://github.com/ahmedalhulaibi/flatfile/tree/master/example) folder in repo for all examples.

Let's say I have a text file `customers.txt` which contains customer data. There are 2 records:

```
AMY1900-01-01019123 FAKE STREETCA
BOB1800-01-01037456 OLD STREET US
```
The file layout is:
```
Columns 1, Length 3 = Customer Name
Columns 4, Length 10 = Customer Join/Open Date
Columns 14, Length 3 = Customer Age
Columns 17, Length 15 = Customer Street Address
Columns 32, Length 2 = Customer Country
```

Each field in a struct can be mapped to a single field in a record using a struct tag.

Struct tags are in the form `flatfile:"col,len"` or for a slice field `flatfile:"col,len,occurrences"`.

I can directly translate these to fields in a struct:

```go
type CustomerRecord struct {
//flatfile is one indexed, column starts at 1
Name string `flatfile:"1,3"`
OpenDate string `flatfile:"4,10"`
Age uint `flatfile:"14,3"`
Address string `flatfile:"17,15"`
CountryCode string `flatfile:"32,2"`
}

```

Once your layout is defined we can read a line from a file and unmarshal the data:

```go
type CustomerRecord struct {
//flatfile is one indexed, column starts at 1
Name string `flatfile:"1,3"`
OpenDate string `flatfile:"4,10"`
Age uint `flatfile:"14,3"`
Address string `flatfile:"17,15"`
CountryCode string `flatfile:"32,2"`
}

func main() {
file, err := os.Open("customers.txt")
checkError(err)
defer file.Close()

reader := bufio.NewReader(file)

eof := false
for !eof {
data := readLine(reader)
if data == nil {
eof = true
} else {
fileRecord := &CustomerRecord{}
//unmarhsal text data to struct
err := flatfile.Unmarshal(data, fileRecord, 0, 0)
fmt.Printf("%v\n", fileRecord)
checkError(err)
}
}
checkError(err)
}

func readLine(reader *bufio.Reader) []byte {
str, _, err := reader.ReadLine()
if err == io.EOF {
return nil
}

return str
}

func checkError(err error) {
if err != nil {
panic(err)
}
}
```

# Features

## Data type support:
- [x] bool
- [x] string
- [x] int
- [x] int8
- [x] int16
- [x] int32
- [x] int64
- [x] uint
- [x] uint8
- [x] uint16
- [x] uint32
- [x] uint64
- [x] float32
- [x] float64
- [x] rune
- [x] byte
- [x] Slice
- [x] Array
- [x] Nested struct
- [x] Nested pointer (to any support type including struct)

- [x] Slice, Array support AKA Emulate [COBOL occurs clause](https://www.ibm.com/support/knowledgecenter/en/SS6SG3_4.2.0/com.ibm.entcobol.doc_4.2/PGandLR/tasks/tptbl03.htm)

- [x] Offset feature to support reading long lines of data. [Example](https://github.com/ahmedalhulaibi/flatfile/tree/master/example/bufferedReadFile)

- [x] Byte and Rune support using type override.

These are aliases for uint8 and int32 respectively. These require an override option to be supplied.
- [x] Flat File abstraction
- [x] Support for conditional unmarshal

if field(col,len) == "text" do unmarshal else skip.
This is done by supplying the `condition` option. The syntax is a bit funky:

`col-len-VALUE` e.g. `1-10-TENLETTERS` This means the unmarshal will only occur if the data from column 1, length 10 is equals to TENLETTERS.

This is useful for flat files where there are multiple record layouts within the same file.