2017-08-05 15 views
0

我必须从数据库中解析10K条目。解析和修复凌乱的字符串数据库的时间

该数据库有一个名为工作小时的字段,显示了汽车经销商的办公室工作时间。

的问题是,这个字段包含描述性的东西是这样的:

Office working hours from 10 am - 4 pm 
Open from 9AM to 5PM 
Main Showroom from 10:00AM - 5:00PM 
Open 10 AM to 13 PM 
Office 10AM to 3PM -- Showroom 9AM to 4PM 

所以你可以看到的风格各不相同,ampm小写和大写,有和没有空间,时间与零和结肠和无它甚至是两种风格产生错误时间的混合。换句话说,一团糟。此外,每行有多个时间范围,或不是。

我想将整个事物转换为24小时格式。

Office working hours from 10:00 - 16:00 hours 
Open from 9:00 to 17:00 hours 
Main Showroom from 10:00 - 17:00 hours 
Open 10:00 to 13:00 hours 
Office 10:00 to 15:00 -- Showroom 9:00 to 16:00 hours 

我可以去的if就像这些无限数量每隔一小时:

if ([text containsString:@"7 PM"]) { 
    text = [text stringByReplacingOccurrencesOfString:@"7 PM" withString:@"19:00"]; 
    } 

,但这将有几十亿行的,不会是有效的。我将不得不测试大写,小写,有无空格和错误的条目。

它必须做一个简单的方法...

任何想法?

+0

结帐NSDateFormatter。 – shallowThought

+0

我认为这是行不通的。这些时间太紧张了。我正在尝试它,它几乎没有捕捉日期。大部分时间都在崩溃。 – SpaceDog

+0

我想知道如果'NSDataDetector'可以帮助。 – Larme

回答

1

NSRegularExpressionNSDateFormatter的组合将会派上用场。结果是这样的:

Results

13PM需要一个人工编辑,或者可以有自动修复的方法。

正则表达式是不完美的,它也将捕获之类的东西36 AM,但日期格式将返回nil

下面的代码,就可以运行:

NSString *test1 = @"Office working hours from 10 am - 4 pm"; 
NSString *test2 = @"Open from 9AM to 5PM"; 
NSString *test3 = @"Main Showroom from 10:00AM - 5:00PM"; 
NSString *test4 = @"Open 10 AM to 13 PM"; 
NSString *test5 = @"Office 10AM to 3PM -- Showroom 9AM to 4PM"; 

NSMutableArray *newStrings = [NSMutableArray array]; 

// [0-9]+ -> Capture 1 or more digit 
// (?:\\:[0-9]+)? -> Capture ":" optionally, if so capture 1 or more digit 
// ()* -> Capture 0 or more whitespace 
// (am|pm) -> Case insensitive search, captures aM, Pm, AM, pm 

NSString *hourPattern = @"([0-9]+(?:\\:[0-9]+)?()*(am|pm))"; 
NSError *error = nil; 
NSRegularExpression *miniFormatter = [NSRegularExpression 
             regularExpressionWithPattern:hourPattern 
             options:NSRegularExpressionCaseInsensitive | NSRegularExpressionSearch 
             error:&error]; 

if(error) 
{ 
    NSLog(@"%@", error.localizedDescription); 
    return; 
} 

for(NSString *text in @[test1, test2, test3, test4, test5]) 
{ 
    NSArray<NSTextCheckingResult *> *matches = [miniFormatter matchesInString:text 
                     options:kNilOptions 
                     range:NSMakeRange(0, text.length)]; 

    NSString *textToChange = [text copy]; 

    for(NSTextCheckingResult *result in matches) 
    { 
     NSString *foundTime = [text substringWithRange:result.range]; 

     NSString *foundTimeOriginal = [foundTime copy]; // This will be used when finding the current range of the text. 

     // Step 1: Remove whitespace for parsing. 

     foundTime = [foundTime stringByReplacingOccurrencesOfString:@" " withString:@""]; 

     // Step 2: Make am/pm uppercase. 

     foundTime = [foundTime uppercaseString]; 

     NSDateFormatter *dateFormatter = [NSDateFormatter new]; 
     [dateFormatter setTimeZone:[NSTimeZone timeZoneWithName:@"GMT"]]; // You may change it accordingly. 

     NSDate *foundDate; 

     // Step 3: Detect if it's in hh:mm format or hh format. 

     if([foundTime containsString:@":"]) 
     { 
      // hh:mm format 

      [dateFormatter setDateFormat:@"hh:mma"]; 
     } 
     else 
     { 
      // hh format 

      [dateFormatter setDateFormat:@"hha"]; 
     } 

     foundDate = [dateFormatter dateFromString:foundTime]; 

     if(!foundDate) 
     { 
      // There's a problem with parsing (such as 13PM). 
      // Proceeding manually... 

      continue; 
     } 

     //NSLog(@"%@ : %@", foundTime, foundDate); 

     // Step 4: Convert to 24-Hour 

     [dateFormatter setDateFormat:@"HH:mm"]; 

     NSString *convertedTime = [dateFormatter stringFromDate:foundDate]; 

     NSRange currentRange = [textToChange rangeOfString:foundTimeOriginal]; 
     textToChange = [textToChange stringByReplacingCharactersInRange:currentRange withString:convertedTime]; 
    } 

    [newStrings addObject:textToChange]; 
} 

for(NSString *text in newStrings) 
{ 
    NSLog(@"%@", text); 
} 

希望这有助于。

+0

哇!辉煌!谢谢!我试图破译你使用过的模式,但我不明白这个部分':\\:' – SpaceDog

+1

不客气! :) 我意识到没有必要''(?:: [0-9] +)?'也可以,但解析是这样的: '?:' - >可选 '\\:' - >逃脱“:” – EDUsta

+1

啊,我明白了。谢谢! – SpaceDog