public void reduce(Text key, Iterable<Text> values, Context context)
throws IOException, InterruptedException {
String dimStr = null;
List<String> playInfoList = new ArrayList<String>();
for (Text val : values) {
String value = val.toString();
String[] valueField = value.split(FIELD_TAB_SEPARATOR, -1);
if (valueField.length == DMInfoHashEnum.MEDIA_NAME.ordinal() + 1) {
dimStr = value;
} else {
playInfoList.add(value);
}
}
for (String playInfoString : playInfoList) {
String playFormatValue = "";
String[] splitMediaPlaySts = playInfoString.split(FIELD_TAB_SEPARATOR, -1);
List<String> splitMediaPlayList = new ArrayList<String>();
for (String splitMediaPlay : splitMediaPlaySts) {
splitMediaPlayList.add(splitMediaPlay);
}
if (null != dimStr) {
String[] dimStrs = dimStr.split(FIELD_TAB_SEPARATOR, -1);
if (dimStrs.length > DMInfoHashEnum.MEDIA_NAME.ordinal()) {
splitMediaPlayList.set(
PlayFormatEnum.SERIAL_ID.ordinal(), dimStrs[DMInfoHashEnum.SERIAL_ID.ordinal()]);
splitMediaPlayList.set(
PlayFormatEnum.MEDIA_ID.ordinal(), dimStrs[DMInfoHashEnum.MEDIA_ID.ordinal()]);
splitMediaPlayList.set(
PlayFormatEnum.CHANNEL_ID.ordinal(), dimStrs[DMInfoHashEnum.CHANNEL_ID.ordinal()]);
splitMediaPlayList.set(
PlayFormatEnum.MEDIA_NAME.ordinal(), dimStrs[DMInfoHashEnum.MEDIA_NAME.ordinal()]);
} else {
splitMediaPlayList.set(PlayFormatEnum.SERIAL_ID.ordinal(), DEFAULT_NEGATIVE_NUM);
splitMediaPlayList.set(PlayFormatEnum.MEDIA_ID.ordinal(), DEFAULT_NEGATIVE_NUM);
}
} else {
if (splitMediaPlayList.size() > PlayFormatEnum.SERIAL_ID.ordinal() + 1) {
splitMediaPlayList.set(PlayFormatEnum.SERIAL_ID.ordinal(), DEFAULT_NEGATIVE_NUM);
splitMediaPlayList.set(PlayFormatEnum.MEDIA_ID.ordinal(), DEFAULT_NEGATIVE_NUM);
}
}
for (int i = 0; i < splitMediaPlayList.size(); i++) {
playFormatValue += splitMediaPlayList.get(i).replaceAll("\\s+", "");
if (i < splitMediaPlayList.size() - 1) {
playFormatValue += FIELD_TAB_SEPARATOR;
}
}
keyText.set(playFormatValue);
context.write(keyText, NullWritable.get());
}
}
public void map(LongWritable key, Text value, Context context)
throws IOException, InterruptedException {
try {
String line = value.toString();
String[] fields = line.split(FIELD_TAB_SEPARATOR, -1);
if (filePath.toLowerCase().contains("play")
&& fields.length > PlayFormatEnum.MEDIA_TYPE_ID.ordinal()) {
String playETLStr = getPlayFormatStr(line);
String[] playField = playETLStr.split(FIELD_TAB_SEPARATOR, -1);
String infohashStr = null;
if (playField[PlayFormatEnum.MEDIA_TYPE_ID.ordinal()].trim().equals("1")
|| playField[PlayFormatEnum.URL.ordinal()].contains("subject/play")) {
infohashStr = playField[PlayFormatEnum.INFOHASH_ID.ordinal()].toUpperCase();
} else {
infohashStr = playField[PlayFormatEnum.MEDIA_ID.ordinal()];
}
if (null != infohashStr && playField.length == PlayFormatEnum.SEIDCOUNT.ordinal() + 1) {
keyText.set(infohashStr.trim());
valueText.set(playETLStr);
context.write(keyText, valueText);
}
} else {
String dimLine = "";
String dimInfo = null;
if (filePath.toLowerCase().contains("infohash")) {
if (fields.length > DMInfoHashEnum.MEDIA_ID.ordinal()) {
dimLine = line.trim();
dimInfo = fields[DMInfoHashEnum.IH.ordinal()];
}
} else if (filePath.toLowerCase().contains("mediainfo")) {
StringBuilder dimStrSb = new StringBuilder();
dimStrSb.append(DEFAULT_INFOHASH + FIELD_TAB_SEPARATOR);
dimStrSb.append(DEFAULT_SERIAL_ID + FIELD_TAB_SEPARATOR);
dimStrSb.append(line.trim());
dimLine = dimStrSb.toString();
dimInfo = fields[DMInfoHashEnum.IH.ordinal()];
}
if (null != dimInfo && !dimInfo.isEmpty()) {
String mediaInfo = dimInfo.trim().toUpperCase();
keyText.set(mediaInfo);
valueText.set(dimLine);
context.write(keyText, valueText);
}
}
} catch (Exception e) {
multipleOutputs.write(
new Text(null == e.getMessage() ? ("error:" + filePath) : e.getMessage()),
new Text(value.toString()),
"_error/part");
e.printStackTrace();
}
}