1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
|
package dstream
import (
"fmt"
"io"
"os"
)
// Dstream streams chunks of data to a consumer.
type Dstream interface {
// Next attempts to advance to the next chunk and returns true
// if successful.
Next() bool
// Names returns the variable names.
Names() []string
// Get returns the values for one variable in the current
// chunk, referring to the variable by name
Get(string) interface{}
// Get returns the values for one variable for the current
// chunk, referring to the variable by position.
GetPos(int) interface{}
// NumVar returns the number of variables in the data set.
NumVar() int
// NumObs returns the number of rows in the data set, it may
// return -1 if not known.
NumObs() int
// Reset sets the provider so that the data are read from the
// beginning of the dataset.
Reset()
// Close frees any resources suh as file handles used by the dstream.
Close()
}
// Dtype represents a data type
type Dtype uint8
// Uint8, etc. are constants defining data types
const (
Uint8 Dtype = iota
Uint16
Uint32
Uint64
Int8
Int16
Int32
Int64
Float32
Float64
Time
String
)
// DataFrame is an implementation of Dstream based on sharded arrays.
type DataFrame struct {
xform // bdata is not used
// data is the underlying data to be passed to the
// consumer.
data [][]interface{}
chunk int // 1-based
done bool
nobs int
}
// Next advances to the next chunk and returns true if successful. If there are
// no more chunks, it returns false.
func (da *DataFrame) Next() bool {
da.chunk++
if da.chunk <= len(da.data[0]) {
return true
}
da.done = true
return false
}
// NumObs returns the number of observations in the DataFrame, if known. If the number
// of observations is not known, it returns -1.
func (da *DataFrame) NumObs() int {
if da.nobs > 0 {
return da.nobs
}
if da.data == nil || len(da.data) == 0 {
// Not yet known
return -1
}
var nobs int
for _, v := range da.data[0] {
nobs += ilen(v)
}
da.nobs = nobs
return nobs
}
func (da *DataFrame) init() {
da.setNamePos() // TODO should get rid of this
}
// Names returns the variable (column) names of the dstream.
func (da *DataFrame) Names() []string {
return da.names
}
// Reset resets the dstream so that after the next call to Next, the
// dstream is at chunk zero.
func (da *DataFrame) Reset() {
da.chunk = 0
da.nobs = 0
da.done = false
}
// GetPos returns the data slice for the variable at the given position.
func (da *DataFrame) GetPos(j int) interface{} {
if da.done {
return nil
}
return da.data[j][da.chunk-1]
}
// Get returns the data slice for the variable with the given name.
func (da *DataFrame) Get(na string) interface{} {
pos := -1
for j, nm := range da.Names() {
if nm == na {
pos = j
break
}
}
if pos == -1 {
msg := fmt.Sprintf("Get: variable '%s' not found", na)
panic(msg)
}
return da.GetPos(pos)
}
// NumVar returns the number of variables in the dstream.
func (da *DataFrame) NumVar() int {
return len(da.data)
}
// NewFromArrays creates a Dstream from raw data stored as slices;
// data[i][j] is the data for the i^th variable in the j^th chunk.
func NewFromArrays(data [][]interface{}, names []string) Dstream {
if len(data) != len(names) {
msg := fmt.Sprintf("NewFromArrays: len(data) = %d != len(names) = %d",
len(data), len(names))
panic(msg)
}
da := &DataFrame{
data: data,
xform: xform{
names: names,
},
}
da.init()
return da
}
// CheckValid runs through the chunks and confirms that the lengths of the slices within
// the chunks are the same. If CheckValid returns false, the dstream is in a corrupted
// state. On completion, the dstream is in its initial state.
func CheckValid(data Dstream) bool {
data.Reset()
names := data.Names()
for c := 0; data.Next(); c++ {
n0 := ilen(data.GetPos(0))
for j := 1; j < len(names); j++ {
n1 := ilen(data.GetPos(j))
if n1 != n0 {
msg := fmt.Sprintf("Length mismatch in chunk %d: len(%s) = %d, len(%s) = %d\n",
c, names[0], n0, names[j], n1)
_, _ = io.WriteString(os.Stderr, msg)
return false
}
}
}
data.Reset()
return true
}
// NewFromFlat creates a Dstream from raw data stored as contiguous
// (flat) arrays. data[i] is the data for the i^th variable, it is
// a slice of fixed-width values, e.g. a []float64.
func NewFromFlat(data []interface{}, names []string) Dstream {
if len(data) != len(names) {
msg := fmt.Sprintf("NewFromFlat: len(data) = %d != len(names) = %d",
len(data), len(names))
panic(msg)
}
var ar [][]interface{}
for _, v := range data {
ar = append(ar, []interface{}{v})
}
da := &DataFrame{
data: ar,
xform: xform{
names: names,
},
}
da.init()
return da
}
// Shallow attempts to make a shallow copy of the data stream.
// Currently, only memory-backed data streams can be shallow copied,
// otherwise a deep copy is returned. Shallow copies of the same data
// can be read in parallel.
func Shallow(data Dstream) Dstream {
data.Reset()
switch data := data.(type) {
case *DataFrame:
var dy DataFrame = *data
return &dy
default:
return MemCopy(data, true)
}
}
|