使用Golang中的go-hdb驱动程序将十进制转换为从SAP HANA数据库检索的字符串

Question

I am new to Golang (but not on HANA), and I am trying to test out the go-hdb to pull out data from TCURR table (from HANA DB SPS6). From what I understand is in that table, UKURS, FFACT, and TFACT are all Decimals in HANA DB. However when retrieving from Go, which I tried fetch using row.Scan to Float64 like this:

var mandt, kurst, fcurr, tcurr, gdatu, datum string
var ukurs float64
var ffact, tfact float64

if err := rows.Scan(&mandt, &kurst, &fcurr, &tcurr, &gdatu, &ukurs, &ffact, &tfact, &datum); err != nil {
   log.Fatal(err)
}

But alas, getting error like this

2016/08/18 10:18:31 sql: Scan error on column index 5: converting driver.Value type []uint8 ("@\xe2\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x0060") to a float64: invalid syntax

I try to read and understand the error it looks like needs uint8 data type (or byte), however, I am stuck with how to convert those uint8/byte into string.

Is there anyone have this problem before and how to getting decimal data and displayed it properly (ie print them into console)?

Answer 1

After to look and understand the driver package a while, I found out there is a function that actually answered my question. The function is reside within decimal.go with function name decodeDecimal(). Due to it is used by internal only, I copy the source code and paste to my code.

The code is something like this:

var mandt, kurst, fcurr, tcurr, gdatu, datum string
var ukurs, ffact, tfact []byte
if err := rows.Scan(&mandt, &kurst, &fcurr, &tcurr, &gdatu, &ukurs, &ffact, &tfact, &datum); err != nil {
   WriteMsg("SCAN")
   log.Fatal(err)
}

var bi big.Int
var z float64
var neg bool
var i int

var record []string

record = append(record, mandt)
record = append(record, kurst)
record = append(record, fcurr)
record = append(record, tcurr)
record = append(record, gdatu)

// ukurs
neg, i = decodeDecimal(ukurs, &bi)
z = BigIntToFloat(neg, &bi, i)
record = append(record, fmt.Sprintf("%.4f", z))

// ffact
neg, i = decodeDecimal(ffact, &bi)
z = BigIntToFloat(neg, &bi, i)
record = append(record, fmt.Sprintf("%.4f", z))

// tfact
neg, i = decodeDecimal(tfact, &bi)
z = BigIntToFloat(neg, &bi, i)
record = append(record, fmt.Sprintf("%.4f", z))

record = append(record, datum)

Function of decodeDecimal()

func decodeDecimal(b []byte, m *big.Int) (bool, int) {

    //bigint word size (*--> src/pkg/math/big/arith.go)
        const (
        dec128Bias = 6176
        // Compute the size _S of a Word in bytes.
        _m    = ^big.Word(0)
        _logS = _m>>8&1 + _m>>16&1 + _m>>32&1
        _S    = 1 << _logS
    )

    neg := (b[15] & 0x80) != 0
    exp := int((((uint16(b[15])<<8)|uint16(b[14]))<<1)>>2) - dec128Bias

    b14 := b[14]  // save b[14]
    b[14] &= 0x01 // keep the mantissa bit (rest: sign and exp)

    //most significand byte
    msb := 14
    for msb > 0 {
        if b[msb] != 0 {
            break
        }
        msb--
    }

    //calc number of words
    numWords := (msb / _S) + 1
    w := make([]big.Word, numWords)

    k := numWords - 1
    d := big.Word(0)
    for i := msb; i >= 0; i-- {
        d |= big.Word(b[i])
        if k*_S == i {
            w[k] = d
            k--
            d = 0
        }
        d <<= 8
    }
    b[14] = b14 // restore b[14]
    m.SetBits(w)
    return neg, exp
}

Function BigIntToFloat()

func BigIntToFloat(sign bool, m *big.Int, exp int) float64 {
    var neg int64
    if sign {
        neg = -1
    } else {
        neg = 1
    }

    return float64(neg*m.Int64()) * math.Pow10(exp)
}