diff --git a/vipra-cmd/src/main/java/de/vipra/cmd/file/FilebaseWordIndex.java b/vipra-cmd/src/main/java/de/vipra/cmd/file/FilebaseWordIndex.java index 51bb48cc593cf1559867d9ef75e6905478fc2237..e035f3ccf988fa6df2225032b7608075fe76aaee 100644 --- a/vipra-cmd/src/main/java/de/vipra/cmd/file/FilebaseWordIndex.java +++ b/vipra-cmd/src/main/java/de/vipra/cmd/file/FilebaseWordIndex.java @@ -53,8 +53,8 @@ public class FilebaseWordIndex implements Iterable<String> { public String transform(final String[] words, final boolean dbInsert) { final CountMap<String> countMap = new CountMap<>(); for (final String word : words) { - if (!word.trim().isEmpty()) { - countMap.count(word); + if (word != null && !word.trim().isEmpty()) { + countMap.count(word.trim()); if (dbInsert) newWords.add(word); } diff --git a/vipra-cmd/src/main/java/de/vipra/cmd/text/ProcessedText.java b/vipra-cmd/src/main/java/de/vipra/cmd/text/ProcessedText.java index 65d93b958cdf0bc9fd946e1474544b57ecc8f1a8..ed941544596b403e8b34b2bd401f2e857e1ca1a8 100644 --- a/vipra-cmd/src/main/java/de/vipra/cmd/text/ProcessedText.java +++ b/vipra-cmd/src/main/java/de/vipra/cmd/text/ProcessedText.java @@ -21,8 +21,8 @@ public class ProcessedText { final String[] allWords = text.toLowerCase().trim().split("\\s+"); final List<String> wordList = new ArrayList<>(allWords.length); for (final String word : allWords) - if (!word.trim().isEmpty()) - wordList.add(word); + if (word != null && !word.trim().isEmpty()) + wordList.add(word.trim()); words = wordList.toArray(new String[allWords.length]); originalWordCount = wordCount; reducedWordCount = words.length;