I am trying to merge two datasets which are firm-level panel data (master) and NAICS to BEA code crosswalk data.
NAICS code in my master data is quite heterogeneous. What is, it varies from 2-digits to 6-digits.
However, NAICS to BEA code crosswalk data looks like the following:
naics | beacode |
111 | 1100 |
112 | 1100 |
113 | 1130 |
114 | 1130 |
115 | 1130 |
211 | 2110 |
212 | 2120 |
213 | 2130 |
22 | 2200 |
23 | 2300 |
311 | 3110 |
312 | 3110 |
313 | 3130 |
314 | 3130 |
315 | 3150 |
316 | 3150 |
321 | 3210 |
322 | 3220 |
323 | 3230 |
324 | 3240 |
325 | 3250 |
326 | 3260 |
327 | 3270 |
331 | 3310 |
332 | 3320 |
333 | 3330 |
334 | 3340 |
335 | 3350 |
336 | 3360 |
337 | 3370 |
339 | 3390 |
42 | 4200 |
44 | 4400 |
45 | 4400 |
481 | 4810 |
482 | 4820 |
483 | 4830 |
484 | 4840 |
485 | 4850 |
486 | 4860 |
487 | 4870 |
488 | 4870 |
492 | 4870 |
493 | 4930 |
511 | 5110 |
512 | 5120 |
513 | 5130 |
514 | 5140 |
515 | 5130 |
516 | 5140 |
517 | 5130 |
518 | 5140 |
519 | 5140 |
521 | 5210 |
522 | 5210 |
523 | 5230 |
524 | 5240 |
525 | 5250 |
531 | 5310 |
532 | 5320 |
533 | 5320 |
5411 | 5411 |
5412 | 5412 |
5413 | 5412 |
5414 | 5412 |
5415 | 5415 |
5416 | 5412 |
5417 | 5412 |
5418 | 5412 |
5419 | 5412 |
55 | 5500 |
561 | 5610 |
562 | 5620 |
61 | 6100 |
621 | 6210 |
622 | 6220 |
623 | 6220 |
624 | 6240 |
711 | 7110 |
712 | 7110 |
713 | 7130 |
721 | 7210 |
722 | 7220 |
81 | 8100 |
What I want to do is, for example, firm A's NAICS code is 817890, however since the crosswalk data does not include this code, I would like to make them as 81 (closest code). Therefore, BEA code for this firm A is 8100.
Or, if I take another example, firm B's NAICS code is 722909. In this case, the closest code in the crosswalk data would be 722, therefore matched BEA code would be 7220.
I hope this example is clear.
In sum, I would like to match my master data with closest NAICS code in the crosswalk data in order to match with BEA code.
Would there be any code to implement this?
Thank you very much in advance,
AC
0 Response to Merge if variable name is similar
Post a Comment